Academic literature on the topic 'Track-before-detect; active sonar tracking'

This work studies the underwater detection and tracking of diver targets under a low signal-to-reverberation ratio (SRR) in active sonar systems. In particular, a particle filter track-before-detect based on a knowledge-aided (KA-PF-TBD) algorithm is proposed. Specifically, the original echo data is directly used as the input of the algorithm, which avoids the information loss caused by threshold detection. Considering the prior motion knowledge of the underwater diver target, we established a multi-directional motion model as the state transition model. An efficient method for calculating the statistical characteristics of echo data about the extended target is proposed based on the non-parametric kernel density estimation theory. The multi-directional movement model set and the statistical characteristics of the echo data are used as the knowledge-aided information of the particle filter process: this is used to calculate the particle weight with the sub-area instead of the whole area, and then the particles with the highest weight are used to estimate the target state. Finally, the effectiveness of the proposed algorithm is proved by simulation and sea-level experimental data analysis through joint evaluation of detection and tracking performance.

The active tracking technology of underwater acoustic targets is an important research direction in the field of underwater acoustic signal processing and sonar, and it has always been issued that draws researchers’ attention. The commonly used Kalman filter active tracking (KFAT) method is an effective tracking method, however, it is difficult to detect weak SNR signals, and it is easy to lose the target after the azimuth of different targets overlaps. This paper proposes a KFAT based on deep convolutional neural network (DCNN) method, which can effectively solve the problem of target loss. First, we use Kalman filtering to predict the azimuth and distance of the target, and then use the trained model to identify the azimuth-weighted time-frequency image to obtain the azimuth and label of the target and obtain the target distance by the time the target appears in the time-frequency image. Finally, we associate the data according to the target category, and update the target azimuth and distance information for this cycle. In this paper, two methods, KFAT and DCNN-KFAT, are simulated and tested, and the results are obtained for two cases of tracking weak signal-to-noise signals and tracking different targets with overlapping azimuths. The simulation results show that the DCNN-KFAT method can solve the problem that the KFAT method is difficult to track the target under the weak SNR and the problem that the target is easily lost when two different targets overlap in azimuth. It reduces the deviation range of the active tracking to within 200 m, which is 500~700 m less than the KFAT method.

Accurate detection and tracking of moving targets in underwater environments pose significant challenges, because noise in acoustic measurements (e.g., SONAR) makes the signal highly stochastic. In continuous marine monitoring a further challenge is related to the computational complexity of the signal processing pipeline—due to energy constraints, in off-shore monitoring platforms algorithms should operate in real time with limited power consumption. In this paper, we present an innovative method that allows to accurately detect and track underwater moving targets from the reflections of an active acoustic emitter. Our system is based on a computationally- and energy-efficient pre-processing stage carried out using a deep convolutional denoising autoencoder (CDA), whose output is then fed to a probabilistic tracking method based on the Viterbi algorithm. The CDA is trained on a large database of more than 20,000 reflection patterns collected during 50 designated sea experiments. System performance is then evaluated on a controlled dataset, for which ground truth information is known, as well as on recordings collected during different sea experiments. Results show that, compared to the benchmark, our method achieves a favorable trade-off between detection and false alarm rate, as well as improved tracking accuracy.

The Ocean Aware project, led by Innovasea and funded through Canada's Ocean Supercluster, is developing a fish passage observation platform to monitor fish without the use of traditional tags. This will provide an alternative to standard tracking technology, such as acoustic telemetry fish tracking, which are often not appropriate for tracking at-risk fish species protected by legislation. Rather, the observation platform uses a combination of sensors including acoustic devices, visual and active sonar, and optical cameras. This will enable more in-depth scientific research and better support regulatory monitoring of at-risk fish species in fish passages or marine energy sites. Analysis of this data will require a robust and accurate method to automatically detect fish, count fish, and classify them by species in real-time using both sonar and optical cameras. To meet this need, we developed and tested an automated real-time deep learning framework combining state of the art convolutional neural networks and Kalman filters. First, we showed that an adaptation of the widely used YOLO machine learning model can accurately detect and classify eight species of fish from a public high resolution DIDSON imaging sonar dataset captured from the Ocqueoc River in Michigan, USA. Although there has been extensive research in the literature identifying particular fish such as eel vs. non-eel and seal vs. fish, to our knowledge this is the first successful application of deep learning for classifying multiple fish species with high resolution imaging sonar. Second, we integrated the Norfair object tracking framework to track and count fish using a public video dataset captured by optical cameras from the Wells Dam fish ladder on the Columbia River in Washington State, USA. Our results demonstrate that deep learning models can indeed be used to detect, classify species, and track fish using both high resolution imaging sonar and underwater video from a fish ladder. This work is a first step toward developing a fully implemented system which can accurately detect, classify and generate insights about fish in a wide variety of fish passage environments and conditions with data collected from multiple types of sensors.

Arrays of tidal turbines are being considered for tidally energetic coastal sites which can be important habitat for many species of marine mammal. Understanding risks to marine mammals from collisions with moving turbine blades must be overcome before regulators can issue licenses for many developments. To understand these risks, it is necessary to understand how animals move around operational turbines and to document the rate at which interactions occur. We report on the design, and performance, of a seabed mounted sensor platform for monitoring the fine scale movements of cetaceans and pinnipeds around operational tidal turbines. The system comprises two high-frequency multibeam active sonars, which can accurately track animals in the horizontal plane. By offsetting the vertical angle of the sonars, the relative intensity of targets on the two sonars can also be used to resolve a vertical component of the animal location. For regularly vocalizing species, i.e., small cetaceans, a tetrahedral array of high frequency hydrophones mounted close to the sonars is used to measure both horizontal and vertical angles to cetacean echolocation clicks. This provides additional localization and tracking information for cetaceans and can also be used to distinguish between pinnipeds and cetaceans detected in the sonar data, based on the presence or absence of echolocation clicks. The system is cabled to shore for power, data transfer, and communications using turbine infrastructure. This allows for continuous operation over many months or years, which will be required to capture what may be rare interactions. The system was tested during a series of multi-week field tests, designed to test system integrity, carry out system calibrations, and test the efficiency of data collection, analyses, and archiving procedures. Overall, the system proved highly reliable, with the PAM system providing bearing accuracies to synthetic sounds of around 4.2 degrees for echolocation clicks with a signal to noise ratio above 15 dB. The system will be deployed close to an operational turbine in early 2022.

Purpose A laser seeker is an important element in missile guidance and control systems, responsible for target detection and tracking. Its control is, however, a challenging problem due to complex dynamics and various acting disturbances. Hence, the purpose of this study is to propose a systematic design, tuning, analysis and performance verification of a nonlinear active disturbance rejection control (ADRC) algorithm for the specific case of the laser seeker system. Design/methodology/approach The proposed systematic approach of nonlinear ADRC application to the laser seeker system consists of the following steps. The complex laser seeker control problem is first expressed as a regulation problem. Then, a nonlinear extended state observer (ESO) with varying gains is used to improve the performance of a conventionally used linear ESO (LESO), which enables better control quality in both transient and steady-state periods. In the next step, a systematic observer tuning, based on a detailed analysis of the system disturbances, is proposed. The stability of the overall control system is then verified using a describing function method. Next, the implementation of the NESO-based ADRC solution is realized in a fixed-point format using MATLAB/Simulink and Xilinx System Generator. Finally, the considered laser seeker control system is implemented in discrete form and comprehensively tested through hardware-in-the-loop (HIL) co-simulation. Findings Through the conducted comparative study of LESO-based and NESO-based ADRC algorithms for the laser seeker system, the advantages of the proposed nonlinear scheme are shown. It is concluded that the NESO-based ADRC scheme for the laser seeker system (with appropriate parameters tuning methodology) provides better control performance in both transient and steady-state periods. The conducted multicriteria study validates the efficacy of the proposed systematic approach of applying nonlinear ADRC to laser seeker systems. Practical implications In practice, the obtained results imply that the laser seeker system, governed by the studied nonlinear version of the ADRC algorithm, could potentially detect and track targets faster and more accurately than the system based on the common linear ADRC algorithm. In addition, the article presents the step-by-step procedure for the design, field programmable gate array (FPGA) implementation and HIL-based co-simulation of the proposed nonlinear controller, which can be used by control practitioners as one of the last validation stages before experimental tests on a real guidance system. Originality/value The main contribution of this work is the systematic procedure of applying the ADRC scheme with NESO for the specific case of the laser seeker system. It includes its design, tuning, analysis and performance verification (with simulation and FPGA hardware). The novelty of the work is also the combination and practical realization of known theoretical elements (NESO structure, NESO parameter tuning, ADRC closed-loop stability analysis) in the specific case of the laser seeker system. The results of the conducted applied research increase the current state of the art related to robust control of laser seeker systems working in disturbed and uncertain conditions.

The use of RFID (radio frequency identification), also called “smart tags”, is on the rise in the retail industry. In short, RFID are tiny microchips using short range radio signals to emit information and can be used to tag goods, buildings, cars, pets, people etc. Unlike bar-code scanners, which must be held directly in front of the item being scanned, one of the benefits of RFID tags is that they can be scanned from a distance. It is expected that RFID will eventually replace the bar code and its use is likely to save companies like Wal-Mart, Procter & Gamble and Gillette millions of dollars as they can track every bottle of shampoo or packet of razor blades from the factory floor to the store shelf (Baard, “Lawmakers”). Most agree that using RFID to track goods from the point of manufacture to the location of sale in order to prevent goods being lost, stolen or handled inappropriately, is acceptable and not cause for privacy concerns. But as marketers often take every opportunity to learn more about consumers and their purchasing behaviours, some fear that tags embedded in clothing, membership cards, mobile phones etc. may be scanned inappropriately and used to target individuals with cleverly tailored marketing messages. In the effort to provide a more customised experience, business is at risk of becoming increasingly intrusive – something that will not be universally acceptable. But is it all bad? Privacy concerns aside, smart tags can add new functions as well as enable a whole range of innovative products and services when joined with other technologies. RFID beyond Traditional Value Chain Management Prada is often mentioned as an example of how RFID can be taken beyond the traditional value chain management. Prada has implemented some ground-breaking technology in their Manhattan (New York) store, all based around RFID. RF-receivers automatically detect and scan garments brought into the dressing room. Via a touch screen the customers view tips on how to mix and match, access information about available sizes, colours, fabrics and styles, and watch video clips of models wearing the very outfit they are trying on (Grassley, ”Prada’s”). Eventually customers will be able to create virtual closets and store information about what they have tried on or bought on their Prada Web account (”Prada’s”). Customers’ details, including notes made by sales assistants, e.g. preferences, can be stored automatically in customer cards, readable by sales assistants’ handhelds or at the cash register (”Prada’s”) – information that could be used by the assistant to spur further sales by suggesting for example: “Last time you were here, you bought a black skirt. We have a sweater that matches that skirt” (Batista). Another example is Precision Dynamics Corp (PDC), which developed an automatic identification wristband incorporating RFID technology. One application is the AgeBand which is used to verify the bearer’s age when purchasing alcohol. ID is checked when entering the venue and the customer receives a plastic wristband printed with personal details that cannot be removed without being damaged or destroyed (Swedberg). The embedded chip can be linked to a customer’s credit card number or a cash deposit to pay for purchases while on the premises. “It is also an easy way to collect statistics for marketing”, says PDC’s senior marketing communications specialist, Paula Maggio (qtd. in Swedberg). Although the RFID clearly provides benefits and new opportunities to business operations, there is an argument over whether consumers will ultimately gain or suffer when smart tags become more commonplace. Certainly it may be convenient to have smart hangers that project virtual clothes onto a customer’s reflection in the mirror so they can try on a range of outfits without having to remove their clothes, but the collection of personal information necessary to provide this convenience also raises complex privacy issues. Fear of Intrusive Marketing Hesseldahl believes that our homes, workplaces, shops, malls, cars, trains, planes and bicycles will all be environments that constantly notice who we are and what we are doing, and which – according to a detailed profile of our habits – will try to service us in ways we can hardly even imagine today (25). This may be helpful to us in many ways, but there are concerns that organisations will use RF-technology to connect product information to individuals in order to create personal profiles which can then be used for pin-pointed marketing purposes, or even for tracking individuals, without their knowledge or consent. A possible scenario is one where consumers are bombarded with intrusive advertising based on what they are wearing, what they are purchasing, their history of past purchases, demographics and more. “Kill Machines” Fearing that the technology will be abused, many privacy advocates suggest that RFID must only be used to keep track of goods in the supply chain and thus be deactivated as soon as they leave the store. For example, consumer organisations such as CASPIAN (Consumers Against Supermarket Privacy Invasion and Numbering), the American Civil Liberties Union, and Electronic Frontiers Australia, lobby for the obligatory deactivation of the tags at the point of purchase. But companies such as Procter & Gamble and Wal-Mart would prefer to keep the tags active after checkout, rather than disabling them with so called “kill machines”, so they can match the unique codes emitted by RF-tags to shopper’s personal information (Baard “Watchdog”). They will want to use RF technology to support the sales process and to provide the consumer with new and better services than what is otherwise possible. And without doubt, if the tags are deactivated some genuinely helpful applications would be lost to the consumer, e.g. being able to call your refrigerator from the supermarket to check if you need milk or your washing machine alerting you if you have accidentally put a delicate garment in your white wash. Looking at the Bright Side Privacy concerns aside, RFID technology, in fact, has the potential to empower consumers as it will put more information about products at consumers’ fingertips. Consumers will for example be able to go into competing supermarkets and scan items with an RF-receiver embedded in their mobile phone, record prices, store and process the information to evaluate which store offers better value. The information can then be shared with other shoppers via the Internet, and suddenly we have a powerful “shopping bot” which transcends the online world. Consequently, RFID has the potential to make competition between retailers tougher than ever before and to benefit consumers through lower and more transparent pricing. In addition, RFID tags may also make possible faster and more accurate services, particularly in supermarkets. Shopping carts are mounted with computers which automatically register all items put into the cart and enable the customer to keep track of items, their prices and their total amount (Blau). RFID can also be used to find the location of items in the store and show more detailed information on a product (origin, use by date, content etc.) and as the customer passes through the checkout, all purchases are registered automatically in a matter of seconds. Privacy Protection In Australia, the Privacy Amendment (Private Sector) Act 2000, with its ten National Privacy Principles (NPP), has been highly criticised over the last few years as being much to open for interpretation and thus difficult to reinforce. In short, the NPP allows for marketers to use non-sensitive personal information for direct marketing purposes without seeking the individual’s consent if it is impracticable to do so (“Guidelines”). That is, as long as they make available a privacy policy explaining why the data is collected and who will have access, they ensure that the data is correct and up to date, protected from unauthorised access, and that individuals are given access to their data upon request (“Guidelines”). In 2003 the Spam Act was introduced in order to take a tougher stand on the escalating problem with massive amounts of unsolicited emails filling up inboxes, threatening the whole concept of the Internet and its many benefits. In essence, the Spam Act will not allow commercial electronic messages to be sent without the recipient’s prior consent or without a possibility to unsubscribe (“Spam”). In the same manner that the Spam Act was passed to regulate the collection of Internet users’ contact information, it may become necessary to regulate the collection and use of data obtained via RFID if the NPP are deemed inadequate. The difficulty will be to do so and at the same time safeguard many of the positive side-effects the technology may have for businesses and consumers. As argued by Roger Clarke, privacy has to be balanced against many other, often competing interests: “The privacy interest of one person may conflict with that of another person, or of a groups of people, or of an organisation, or of society as a whole. Hence: Privacy Protection is a process of finding appropriate balances between privacy and multiple competing interests.” It is therefore recommended that legislators and policy makers keep up with the development and undertake significant research into both sides before any legislation is passed so that the best interests of consumers and business are catered for. Can There Be a Win-Win Situation? Although business can expect some significant gains from the use of RFID, particularly through a more effective value chain management but also from more substantial and better quality business intelligence, consumers may in fact be the real winners as new and better business concepts, products and services are made available. Further, with the increased transparency in business, consumers can use the vast amounts of information available to find the best products and services, at the best price, and from the best provider. With the aid of smart software, such as search agents, it will be a rather effortless task and will provide consumers with a real advantage. But this assumes consumers are aware of the benefits and how they can be exploited, and have the means to do so – something that will require some skill, interest, money and time. Consumers will also have to give up some privacy in order to take full advantage of the new technology. For the industry, the main challenge will be communicating what these advantages are, as acceptance, adoption and thus also return on investment will depend upon it. For legislators and policy makers, the major dilemma will be to provide a regulatory framework that is flexible but distinct, and that will prevent abuse and at the same time enable positive outcomes for both business and consumers. A fine line that should be treaded wisely in order to create a future where everyone can gain from the benefits of using this technology. References Baard, Mark. “Lawmakers Alarmed by RFID Spying.” Wired News 26 Feb. 2004. 9 Mar. 2004 http://www.wired.com/news/privacy/0,1848,62433,00.html>. Baard, Mark. ”Watchdog Push for RFID Laws.” Wired News 5 Apr. 2004. 6 Apr. 2004 http://www.wired.com/news/privacy/0,1848,62922,00.html>. Batista, Elisa. ”What Your Clothes Say about You.” Wired News 12 Mar. 2003. 8 Mar. 2004 http://www.wired.com/news/wireless/0,1382,58006,00.html>. Blau, John. ”Så fungerar det digitala snabbköpet.” PC för Alla 1 (2004). 8 Mar. 2004 http://www.idg.se/ArticlePages/200402/27/20040227165630_IDG. se760/20040227165630_IDG.se760.dbp.asp>. Clarke, Roger. Beyond the OECD Guidelines: Privacy Protection for the 21st Century. 4 Jan. 2000. 15 Mar. 2004 http://www.anu.edu.au/people/Roger.Clarke/DV/PP21C.html>. Grassley, Tanya. ”Retailers Outfit Stores with Tech.” Wired News 18 Dec. 2002. 8 Mar. 2004 http://www.wired.com/news/holidays/0,1882,56885,00.html>. “Guidelines to the National Privacy Principles.” The Office of the Federal Privacy Commissioner 2001. 4 Apr. 2004 http://www.privacy.gov.au/publications/nppgl_01.html#sum>. Hessledahl, Peter. Den globale organisme. Copenhagen: Aschehoug, 2002. 24 Apr. 2004 http://www.global-organisme.dk/e-bog/den_globale_organisme.pdf>. ”Prada’s Smart Tags Too Clever?” Wired News 27 Oct. 2002. 9 Mar. 2004 http://www.wired.com/news/technology/0,1282,56042,00.html>. “Spam” DCITA 2004. 4 Apr. 2004 http://www2.dcita.gov.au/ie/trust/improving/spam>. Swedberg, Claire. ”Putting Drinks on the Cuff.” RFID Journal 15 Jun. 2004. 15 Jun. 2004 http://www.rfidjournal.com/article/articleview/987/1/1/>. Citation reference for this article MLA Style Melchior, Angelika. "Tag & Trace Marketing." M/C Journal 8.4 (2005). echo date('d M. Y'); ?> <http://journal.media-culture.org.au/0508/03-melchior.php>. APA Style Melchior, A. (Aug. 2005) "Tag & Trace Marketing," M/C Journal, 8(4). Retrieved echo date('d M. Y'); ?> from <http://journal.media-culture.org.au/0508/03-melchior.php>.

Introduction The accelerometer seems, at first, both advanced and dated, both too complex and not complex enough. It sits in our video game controllers and our smartphones allowing us to move beyond mere button presses into immersive experiences where the motion of the hand is directly translated into the motion on the screen, where our flesh is transformed into the flesh of a superhero. Or at least that was the promise in 2005. Since then, motion control has moved from a promised revitalization of the video game industry to a not-quite-good-enough gimmick that all games use but none use well. Rogers describes the diffusion of innovation, as an invention or technology comes to market, in five phases: First, innovators will take risks with a new invention. Second, early adopters will establish a market and lead opinion. Third, the early majority shows that the product has wide appeal and application. Fourth, the late majority adopt the technology only after their skepticism has been allayed. Finally the laggards adopt the technology only when no other options are present (62). Not every technology makes it through the diffusion, however, and there are many who have never warmed to the accelerometer-controlled video game. Once an innovation has moved into the mainstream, additional waves of innovation may take place, when innovators or early adopters may find new uses for existing technology, and bring these uses into the majority. This is the case with the accelerometer that began as an airbag trigger and today is used for measuring and augmenting human motion, from dance to health (Walter 84). In many ways, gestural control of video games, an augmentation technology, was an interlude in the advancement of motion control. History In the early 1920s, bulky proofs-of-concept were produced that manipulated electrical voltage levels based on the movement of a probe, many related to early pressure or force sensors. The relationships between pressure, force, velocity and acceleration are well understood, but development of a tool that could measure one and infer the others was a many-fronted activity. Each of these individual sensors has its own specific application and many are still in use today, as pressure triggers, reaction devices, or other sensor-based interactivity, such as video games (Latulipe et al. 2995) and dance (Chu et al. 184). Over the years, the probes and devices became smaller and more accurate, and eventually migrated to the semiconductor, allowing the measurement of acceleration to take place within an almost inconsequential form-factor. Today, accelerometer chips are in many consumer devices and athletes wear battery-powered wireless accelerometer bracelets that report their every movement in real-time, a concept unimaginable only 20 years ago. One of the significant initial uses for accelerometers was as a sensor for the deployment of airbags in automobiles (Varat and Husher 1). The sensor was placed in the front bumper, detecting quick changes in speed that would indicate a crash. The system was a significant advance in the safety of automobiles, and followed Rogers’ diffusion through to the point where all new cars have airbags as a standard component. Airbags, and the accelerometers which allow them to function fast enough to save lives, are a ubiquitous, commoditized technology that most people take for granted, and served as the primary motivating factor for the mass-production of silicon-based accelerometer chips. On 14 September 2005, a device was introduced which would fundamentally alter the principal market for accelerometer microchips. The accelerometer was the ADXL335, a small, low-power, 3-Axis device capable of measuring up to 3g (1g is the acceleration due to gravity), and the device that used this accelerometer was the Wii remote, also called the Wiimote. Developed by Nintendo and its holding companies, the Wii remote was to be a defining feature of Nintendo’s 7th-generation video game console, in direct competition with the Xbox 360 and the Playstation 3. The Wii remote was so successful that both Microsoft and Sony added motion control to their platforms, in the form of the accelerometer-based “dual shock” controller for the Playstation, and later the Playstation Move controller; as well as an integrated accelerometer in the Xbox 360 controller and the later release of the Microsoft Kinect 3D motion sensing camera. Simultaneously, computer manufacturing companies saw a different, more pedantic use of the accelerometer. The primary storage medium in most computers today is the Hard Disk Drive (HDD), a set of spinning platters of electro-magnetically stored information. Much like a record player, the HDD contains a “head” which sweeps back and forth across the platter, reading and writing data. As computers changed from desktops to laptops, people moved their computers more often, and a problem arose. If the HDD inside a laptop was active when the laptop was moved, the read head might touch the surface of the disk, damaging the HDD and destroying information. Two solutions were implemented: vibration dampening in the manufacturing process, and the use of an accelerometer to detect motion. When the laptop is bumped, or dropped, the hard disk will sense the motion and immediately park the head, saving the disk and the valuable data inside. As a consequence of laptop computers and Wii remotes using accelerometers, the market for these devices began to swing from their use within car airbag systems toward their use in computer systems. And with an accelerometer in every computer, it wasn’t long before clever programmers began to make use of the information coming from the accelerometer for more than just protecting the hard drive. Programs began to appear that would use the accelerometer within a laptop to “lock” it when the user was away, invoking a loud noise like a car alarm to alert passers-by to any potential theft. Other programmers began to use the accelerometer as a gaming input, and this was the beginning of gesture control and the augmentation of human motion. Like laptops, most smartphones and tablets today have accelerometers included among their sensor suite (Brezmes et al. 796). These accelerometers strictly a user-interface tool, allowing the phone to re-orient its interface based on how the user is holding it, and allowing the user to play games and track health information using the phone. Many other consumer electronic devices use accelerometers, such as digital cameras for image stabilization and landscape/portrait orientation. Allowing a device to know its relative orientation and motion provides a wide range of augmentation possibilities. The Language of Measuring Motion When studying accelerometers, their function, and applications, a critical first step is to examine the language used to describe these devices. As the name implies, the accelerometer is a device which measures acceleration, however, our everyday connotation of this term is problematic at best. In colloquial language, we say “accelerate” when we mean “speed up”, but this is, in fact, two connotations removed from the physical property being measured by the device, and we must unwrap these layers of meaning before we can understand what is being measured. Physicists use the term “accelerate” to mean any change in velocity. It is worth reminding ourselves that velocity (to the physicists) is actually a pair of quantities: a speed coupled with a direction. Given this definition, when an object changes velocity (accelerates), it can be changing its speed, its direction, or both. So a car can be said to be accelerating when speeding up, slowing down, or even turning while maintaining a speed. This is why the accelerometer could be used as an airbag sensor in the first place. The airbags should deploy when a car suddenly changes velocity in any direction, including getting faster (due to being hit from behind), getting slower (from a front impact crash) or changing direction (being hit from the side). It is because of this ability to measure changes in velocity that accelerometers have come into common usage for laptop drop sensors and video game motion controllers. But even this understanding of accelerometers is incomplete. Because of the way that accelerometers are constructed, they actually measure “proper acceleration” within the context of a relativistic frame of reference. Discussing general relativity is beyond the scope of this paper, but it is sufficient to describe a relativistic frame of reference as one in which no forces are felt. A familiar example is being in orbit around the planet, when astronauts (and their equipment) float freely in space. A state of “free-fall” is one in which no forces are felt, and this is the only situation in which an accelerometer reads 0 acceleration. Since most of us are not in free-fall most of the time, any accelerometers in devices in normal use do not experience 0 proper acceleration, even when apparently sitting still. This is, of course, because of the force due to gravity. An accelerometer sitting on a table experiences 1g of force from the table, acting against the gravitational acceleration. This non-zero reading for a stationary object is the reason that accelerometers can serve a second (and, today, much more common) use: measuring orientation with respect to gravity. Gravity and Tilt Accelerometers typically measure forces with respect to three linear dimensions, labeled x, y, and z. These three directions orient along the axes of the accelerometer chip itself, with x and y normally orienting along the long faces of the device, and the z direction often pointing through the face of the device. Relative motion within a gravity field can easily be inferred assuming that the only force acting on the device is gravity. In this case, the single force is distributed among the three axes depending on the orientation of the device. This is how personal smartphones and video game controllers are able to use “tilt” control. When held in a natural position, the software extracts the relative value on all three axes and uses that as a reference point. When the user tilts the device, the new direction of the gravitational acceleration is then compared to the reference value and used to infer the tilt. This can be done hundreds of times a second and can be used to control and augment any aspect of the user experience. If, however, gravity is not the only force present, it becomes more difficult to infer orientation. Another common use for accelerometers is to measure physical activity like walking steps. In this case, it is the forces on the accelerometer from each footfall that are interpreted to measure fitness features. Tilt is unreliable in this circumstance because both gravity and the forces from the footfall are measured by the accelerometer, and it is impossible to separate the two forces from a single measurement. Velocity and Position A second common assumption with accelerometers is that since they can measure acceleration (rate of change of velocity), it should be possible to infer the velocity. If the device begins at rest, then any measured acceleration can be interpreted as changes to the velocity in some direction, thus inferring the new velocity. Although this is theoretically possible, real-world factors come in to play which prevent this from being realized. First, the assumption of beginning from a state of rest is not always reasonable. Further, if we don’t know whether the device is moving or not, knowing its acceleration at any moment will not help us to determine it’s new speed or position. The most important real-world problem, however, is that accelerometers typically show small variations even when the object is at rest. This is because of inaccuracies in the way that the accelerometer itself is interpreted. In normal operation, these small changes are ignored, but when trying to infer velocity or position, these little errors will quickly add up to the point where any inferred velocity or position would be unreliable. A common solution to these problems is in the combination of devices. Many new smartphones combine an accelerometer and a gyroscopes (a device which measures changes in rotational inertia) to provide a sensing system known as an IMU (Inertial measurement unit), which makes the readings from each more reliable. In this case, the gyroscope can be used to directly measure tilt (instead of inferring it from gravity) and this tilt information can be subtracted from the accelerometer reading to separate out the motion of the device from the force of gravity. Augmentation Applications in Health, Gaming, and Art Accelerometer-based devices have been used extensively in healthcare (Ward et al. 582), either using the accelerometer within a smartphone worn in the pocket (Yoshioka et al. 502) or using a standalone accelerometer device such as a wristband or shoe tab (Paradiso and Hu 165). In many cases, these devices have been used to measure specific activity such as swimming, gait (Henriksen et al. 288), and muscular activity (Thompson and Bemben 897), as well as general activity for tracking health (Troiano et al. 181), both in children (Stone et al. 136) and the elderly (Davis and Fox 581). These simple measurements are the first step in allowing athletes to modify their performance based on past activity. In the past, athletes would pour over recorded video to analyze and improve their performance, but with accelerometer devices, they can receive feedback in real time and modify their own behaviour based on these measurements. This augmentation is a competitive advantage but could be seen as unfair considering the current non-equal access to computer and electronic technology, i.e. the digital divide (Buente and Robbin 1743). When video games were augmented with motion controls, many assumed that this would have a positive impact on health. Physical activity in children is a common concern (Treuth et al. 1259), and there was a hope that if children had to move to play games, an activity that used to be considered a problem for health could be turned into an opportunity (Mellecker et al. 343). Unfortunately, the impact of children playing motion controlled video games has been less than successful. Although fitness games have been created, it is relatively easy to figure out how to activate controls with the least possible motion, thereby nullifying any potential benefit. One of the most interesting applications of accelerometers, in the context of this paper, is the application to dance-based video games (Brezmes et al. 796). In these systems, participants wear devices originally intended for health tracking in order to increase the sensitivity and control options for dance. This has evolved both from the use of accelerometers for gestural control in video games and for measuring and augmenting sport. Researchers and artists have also recently used accelerometers to augment dance systems in many ways (Latulipe et al. 2995) including combining multiple sensors (Yang et al. 121), as discussed above. Conclusions Although more and more people are using accelerometers in their research and art practice, it is significant that there is a lack of widespread knowledge about how the devices actually work. This can be seen in the many art installations and sports research studies that do not take full advantage of the capabilities of the accelerometer, or infer information or data that is unreliable because of the way that accelerometers behave. This lack of understanding of accelerometers also serves to limit the increased utilization of this powerful device, specifically in the context of augmentation tools. Being able to detect, analyze and interpret the motion of a body part has significant applications in augmentation that are only starting to be realized. The history of accelerometers is interesting and varied, and it is worthwhile, when exploring new ideas for applications of accelerometers, to be fully aware of the previous uses, current trends and technical limitations. It is clear that applications of accelerometers to the measurement of human motion are increasing, and that many new opportunities exist, especially in the application of combinations of sensors and new software techniques. The real novelty, however, will come from researchers and artists using accelerometers and sensors in novel and unusual ways. References Brezmes, Tomas, Juan-Luis Gorricho, and Josep Cotrina. “Activity Recognition from Accelerometer Data on a Mobile Phone.” In Distributed Computing, Artificial Intelligence, Bioinformatics, Soft Computing, and Ambient Assisted Living. Springer, 2009. Buente, Wayne, and Alice Robbin. “Trends in Internet Information Behavior, 2000-2004.” Journal of the American Society for Information Science and Technology 59.11 (2008).Chu, Narisa N.Y., Chang-Ming Yang, and Chih-Chung Wu. “Game Interface Using Digital Textile Sensors, Accelerometer and Gyroscope.” IEEE Transactions on Consumer Electronics 58.2 (2012): 184-189. Davis, Mark G., and Kenneth R. Fox. “Physical Activity Patterns Assessed by Accelerometry in Older People.” European Journal of Applied Physiology 100.5 (2007): 581-589.Hagstromer, Maria, Pekka Oja, and Michael Sjostrom. “Physical Activity and Inactivity in an Adult Population Assessed by Accelerometry.” Medical Science and Sports Exercise. 39.9 (2007): 1502-08. Henriksen, Marius, H. Lund, R. Moe-Nilssen, H. Bliddal, and B. Danneskiod-Samsøe. “Test–Retest Reliability of Trunk Accelerometric Gait Analysis.” Gait & Posture 19.3 (2004): 288-297. Latulipe, Celine, David Wilson, Sybil Huskey, Melissa Word, Arthur Carroll, Erin Carroll, Berto Gonzalez, Vikash Singh, Mike Wirth, and Danielle Lottridge. “Exploring the Design Space in Technology-Augmented Dance.” In CHI’10 Extended Abstracts on Human Factors in Computing Systems. ACM, 2010. Mellecker, Robin R., Lorraine Lanningham-Foster, James A. Levine, and Alison M. McManus. “Energy Intake during Activity Enhanced Video Game Play.” Appetite 55.2 (2010): 343-347. Paradiso, Joseph A., and Eric Hu. “Expressive Footwear for Computer-Augmented Dance Performance.” In First International Symposium on Wearable Computers. IEEE, 1997. Rogers, Everett M. Diffusion of Innovations. New York: Free Press of Glencoe, 1962. Stone, Michelle R., Ann V. Rowlands, and Roger G. Eston. "Relationships between Accelerometer-Assessed Physical Activity and Health in Children: Impact of the Activity-Intensity Classification Method" The Free Library 1 Mar. 2009. Thompson, Christian J., and Michael G. Bemben. “Reliability and Comparability of the Accelerometer as a Measure of Muscular Power.” Medicine and Science in Sports and Exercise. 31.6 (1999): 897-902.Treuth, Margarita S., Kathryn Schmitz, Diane J. Catellier, Robert G. McMurray, David M. Murray, M. Joao Almeida, Scott Going, James E. Norman, and Russell Pate. “Defining Accelerometer Thresholds for Activity Intensities in Adolescent Girls.” Medicine and Science in Sports and Exercise 36.7 (2004):1259-1266Troiano, Richard P., David Berrigan, Kevin W. Dodd, Louise C. Masse, Timothy Tilert, Margaret McDowell, et al. “Physical Activity in the United States Measured by Accelerometer.” Medicine and Science in Sports and Exercise, 40.1 (2008):181-88. Varat, Michael S., and Stein E. Husher. “Vehicle Impact Response Analysis through the Use of Accelerometer Data.” In SAE World Congress, 2000. Walter, Patrick L. “The History of the Accelerometer”. Sound and Vibration (Mar. 1997): 16-22. Ward, Dianne S., Kelly R. Evenson, Amber Vaughn, Anne Brown Rodgers, Richard P. Troiano, et al. “Accelerometer Use in Physical Activity: Best Practices and Research Recommendations.” Medicine and Science in Sports and Exercise 37.11 (2005): S582-8. Yang, Chang-Ming, Jwu-Sheng Hu, Ching-Wen Yang, Chih-Chung Wu, and Narisa Chu. “Dancing Game by Digital Textile Sensor, Accelerometer and Gyroscope.” In IEEE International Games Innovation Conference. IEEE, 2011.Yoshioka, M., M. Ayabe, T. Yahiro, H. Higuchi, Y. Higaki, J. St-Amand, H. Miyazaki, Y. Yoshitake, M. Shindo, and H. Tanaka. “Long-Period Accelerometer Monitoring Shows the Role of Physical Activity in Overweight and Obesity.” International Journal of Obesity 29.5 (2005): 502-508.

Detection of an underwater target with active sonar in shallow waters such as the Baltic sea is a big challenge. This since the sound beams from the sonar will be reflected on the surfaces, sea surface and sea bottom, and the water volume itself which generates reverberation. Reverberation which will be reflected back to the receiver, is strong in intensity which give rise to many false targets in terms of classifying a target in a surveillance area. These false targets are unwanted and a real target might benefit from these miss-classifications in terms of remaining undetected. It is especially hard if the signal-to-noise ratio (SNR) is approaching zero, i.e. the target strength and the reverberation strength are equal in magnitude. The classical approach to a target detection problem is to assign a threshold value to the measurement, and the data point exceeding the threshold is classified as a target. This approach does not hold for low levels of SNR, since a threshold would not have a statistical significance and could lead to neglecting important data. Track-before-detect (TrBD) is a proposed method for low-SNR situations which tracks and detects a target based on unthresholded data. TrBD enables tracking and detecting of weak and/or stealthy targets. Due to the issues with target detection in shallow waters, the hypothesis of this thesis is to investigate the possibility to implement TrBD, and evaluate the performance of it, when applied on a low-SNR target. The TrBD is implemented with a particle filter which is a recursive Bayesian solution to the problem of integrated tracking and detection. The reverberation data was generated by filtering white noise with an Autoregressive filter of order 1. The target is assigned to propagate according to a constant velocity state space model. Two types of TrBD algorithms are implemented, one which is trained on the background and one which is not. The untrained TrBD is able to track and detect the target but only for levels of SNR down to 4dB. Lower SNR leads to the algorithm not being able to distinguish the target signal from the reverberation. The trained TrBD on the other hand, is able to perform very well for levels of SNR down to 0dB, it is able to track and detect the target and neglect the reverberation. For trajectories passing through areas with high reverberation, the target was lost for a short period of time until it could be retracked again. Overall, the TrBD was successfully implemented on the self-generated data and has a good performance for various target trajectories.

Detection and tracking of signals used in sonar applications in noisy environment is the focus of this thesis. We have concentrated on the low Signal-to-Noise Ratio (SNR) case where the conventional detection methods are not applicable. Furthermore, it is assumed that the duty cycle is relatively low. In the problem that is of concern the carrier frequency, pulse repetition interval (PRI) and the existence of the signal are not known. The unknown character of PRI makes the problem challenging since it means that the signal exists at some unknown intervals. A recursive, Bayesian track-before-detect (TBD) filter using particle filter based methods is proposed to solve the concerned problem. The data used by the particle filter is the magnitude of a complex spectrum in complex Gaussian noise. The existence variable is added in the design of the filter to determine the existence of the signal. The evolution of the signal state is modeled by a linear stochastic process. The filter estimates the signal state including the carrier frequency and PRI. Simulations are done under different scenarios where the carrier frequency, PRI and the existence of the signal varies. The results demonstrate that the algorithm presented in this thesis can detect signals which cannot be detected by conventional methods. Besides detection, the tracking performance of the filter is satisfying.

The detection and tracking of underwater targets with active sonar is a challenging problem because of high acoustic clutter, fluctuating target returns and a relatively low measurement update rate. In this thesis, a Bayesian framework for the detection and tracking of underwater targets using active sonar is formulated. In general, Bayesian tracking algorithms are built on two statistical models: the target dynamics model and the measurement model. The target dynamics model describes the evolution of the target state with time and is almost always assumed to be a Markov process. The typical measurement model approximates the sensor image with a collection of discrete points at each frame and allows point measurement tracking to be performed. This thesis investigates alternative target and measurement models and considers their application to active sonar tracking. The Markov process commonly used for target modelling assumes that the state evolves without knowledge of its future destination. Random realisations of a Markov process can also display a large amount of variability and do not, in general, resemble realistic target trajectories. An alternative is the reciprocal process, which assumes conditioning on a known destination state. The first key contribution is the derivation and implementation of a Maximum Likelihood Sequence Estimator (MLSE) for a Hidden Reciprocal Process (HRP). The performance of the proposed algorithm is demonstrated in simulated scenarios and shown to give improved state estimation performance over Markov processes for scenarios featuring reciprocal targets. In point measurement tracking, reducing the sensor data to point detections results in the loss of valuable information. This method is generally sufficient for tracking high Signal-to-Noise Ratio (SNR) targets but can fail in the case of low SNR targets. The alternative to point measurement tracking is to provide the sensor intensity map, an image, as an input into the tracker. This paradigm is referred to as Track-Before-Detect (TkBD). This thesis will focus on a particular TkBD algorithm based on Expectation-Maximisation (EM) data association called the Histogram-Probabilistic Multi-Hypothesis Tracker (H-PMHT) as it handles multiple targets with low complexity. In the second key contribution, we demonstrate a Viterbi implementation of the H-PMHT algorithm, and show that it outperforms the Kalman Filter in the linear non-Gaussian case. A problem with H-PMHT is that it fails to model fluctuating target amplitude, which can degrade performance in realistic sensing conditions. The third key contribution addresses this by replacing the multinomial measurement model with a Poisson mixture process. The new Poisson mixture is shown to be consistent with the original H-PMHT modelling assumptions but it now allows for a randomly evolving mean target amplitude state with instantaneous fluctuations. This new TkBD algorithm is referred to as the Poisson H-PMHT. The Bayesian prior on the target state is also modified to ensure more robust performance. The fourth contribution is a novel TkBD algorithm based on the application of EM data association to a new measurement model that directly describes continuous valued intensity maps and avoids using an intermediate quantisation stage like the H-PMHT. This model is referred to as the Interpolated Poisson measurement model and is integrated into the Probabilistic Multi-Hypothesis Tracker (PMHT) framework to derive a TkBD algorithm for continuous data called the Interpolated Poisson-PMHT (IP-PMHT). The performance of the Poisson H-PMHT and IP-PMHT algorithms are verified through simulations and are shown to outperform the standard H-PMHT in terms of SNR estimation, particularly for scenarios featuring targets with highly fluctuating amplitude. The final key contribution is the application of several TkBD algorithms based on EM data association to the active sonar problem through a comparative study using trial data from an active towed array sonar. The TkBD algorithms are modified to incorporate changes in target appearance with received array bearing, and are shown to give improved SNR and state estimation performance compared with a conventional point measurement tracking algorithm. The thesis concludes by discussing the limitations of the proposed algorithms and possible avenues for future work.
Thesis (Ph.D.) -- University of Adelaide, School of Electrical and Electronic Engineering, 2015