No description available.

Chemical sensors are generally used as one-dimensional devices, where one measures the sensor’s response at a fixed setting, e.g., infrared absorption at a specific wavelength, or conductivity of a solid-state sensor at a specific operating temperature. In many cases, additional information can be extracted by modulating some internal property (e.g., temperature, voltage) of the sensor. However, this additional information comes at a cost (e.g., sensing times, power consumption), so offline optimization techniques (such as feature-subset selection) are commonly used to identify a subset of the most informative sensor tunings. An alternative to offline techniques is active sensing, where the sensor tunings are adapted in real-time based on the information obtained from previous measurements. Prior work in domains such as vision, robotics, and target tracking has shown that active sensing can schedule agile sensors to manage their sensing resources more efficiently than passive sensing, and also balance between sensing costs and performance. Inspired from the history of active sensing, in this dissertation, we developed active sensing algorithms that address three different computational problems in chemical sensing. First, we consider the problem of classification with a single tunable chemical sensor. We formulate the classification problem as a partially observable Markov decision process, and solve it with a myopic algorithm. At each step, the algorithm estimates the utility of each sensing configuration as the difference between expected reduction in Bayesian risk and sensing cost, and selects the configuration with maximum utility. We evaluated this approach on simulated Fabry-Perot interferometers (FPI), and experimentally validated on metal-oxide (MOX) sensors. Our results show that the active sensing method obtains better classification performance than passive sensing methods, and also is more robust to additive Gaussian noise in sensor measurements. Second, we consider the problem of estimating concentrations of the constituents in a gas mixture using a tunable sensor. We formulate this multicomponent-analysis problem as that of probabilistic state estimation, where each state represents a different concentration profile. We maintain a belief distribution that assigns a probability to each profile, and update the distribution by incorporating the latest sensor measurements. To select the sensor’s next operating configuration, we use a myopic algorithm that chooses the operating configuration expected to best reduce the uncertainty in the future belief distribution. We validated this approach on both simulated and real MOX sensors. The results again demonstrate improved estimation performance and robustness to noise. Lastly, we present an algorithm that extends active sensing to sensor arrays. This algorithm borrows concepts from feature subset selection to enable an array of tunable sensors operate collaboratively for the classification of gas samples. The algorithm constructs an optimized action vector at each sensing step, which contains separate operating configurations for each sensor in the array. When dealing with sensor arrays, one needs to account for the correlation among sensors. To this end, we developed two objective functions: weighted Fisher scores, and dynamic mutual information, which can quantify the discriminatory information and redundancy of a given action vector with respect to the measurements already acquired. Once again, we validated the approach on simulated FPI arrays and experimentally tested it on an array of MOX sensors. The results show improved classification performance and robustness to additive noise.

Context information on a smart phone can be used to tailor applications for specific situations (e.g. provide tailored routing advice based on location, gas prices and traffic). However, typical context-aware smart phone applications use very limited context information such as user identity, location and time. In the future, smart phones will need to decide from a wide range of sensors to gather information from in order to best accommodate user needs and preferences in a given context. In this thesis, we present a model for active sensor selection within decision-making processes, in which observational features are selected based on longer-term impact on the decisions made by the smart phone. This thesis formulates the problem as a partially observable Markov decision process (POMDP), and proposes a non-myopic solution to the problem using a state of the art approximate planning algorithm Symbolic Perseus. We have tested our method on a 3 small example domains, comparing different policy types, discount factors and cost settings. The experimental results proved that the proposed approach delivers a better policy in the situation of costly sensors, while at the same time provides the advantage of faster policy computation with less memory usage.

Context information on a smart phone can be used to tailor applications for specific situations (e.g. provide tailored routing advice based on location, gas prices and traffic). However, typical context-aware smart phone applications use very limited context information such as user identity, location and time. In the future, smart phones will need to decide from a wide range of sensors to gather information from in order to best accommodate user needs and preferences in a given context. In this thesis, we present a model for active sensor selection within decision-making processes, in which observational features are selected based on longer-term impact on the decisions made by the smart phone. This thesis formulates the problem as a partially observable Markov decision process (POMDP), and proposes a non-myopic solution to the problem using a state of the art approximate planning algorithm Symbolic Perseus. We have tested our method on a 3 small example domains, comparing different policy types, discount factors and cost settings. The experimental results proved that the proposed approach delivers a better policy in the situation of costly sensors, while at the same time provides the advantage of faster policy computation with less memory usage.

The increasing use of piezoelectric material in the last decade has led to the new discovery that piezoelectric devices may be used not only as sensors or actuators individually, but also as a sensor and actuator at the same time. This bi-directional phenomenon greatly expands the utility of piezoelectric devices, especially in the area of smart material systems and structures, such as structural dynamic measurement analysis, structural damage detection, active structural vibration and acoustic control, etc. Presented in this thesis is a new electromechanical approach for structural dynamic analysis. It uses PZT patches bonded on structures as active piezoelectric sensors, and acquires the structural information by measuring the electric admittance of the sensors. Because of the electromechanical coupling, the electric admittance is mechanically modulated by the structural dynamics through the piezoelectric effect. The structural dynamic characteristics can then be extracted using a mathematical model governing the interaction of PZT actuators and structures. In this thesis, a multi-input, multi-output mathematical model is derived based on the one-dimensional constitutive law of the PZT and the mechanical impedance approach. The model yields explicit expressions of structural mechanical mobility in terms of the measured electric admittance. Applications of the approach in smart structures are presented in the area of mechanical frequency response function extraction, structural modal analysis, and structural health monitoring. As a general issue related to the application of PZT sensor-actuators in smart structures, the energy conversion efficiency of PZT actuators has also been experimentally investigated. / Master of Science

Animals must sense their surroundings to update their internal representations of the external environment, and exploratory behaviours such as sensory sampling are influenced by past experiences. This thesis investigates how voluntary sensory sampling activities undergo learning-dependent changes. Studies of freely behaving animals impose two major challenges: 1) the accuracy of biological measurements is compromised by movement-induced artifacts, and 2) large degrees of freedom in unrestrained behaviours confound well-controlled studies. Pulse-type weakly electric fish (WEF) are an ideal choice to study adaptive sensory sampling from unrestrained animals, since they generate readily observable and quantifiable sensory capture events expressed by discrete pulses of electric organ discharges (EODs). To study the voluntarily movements and sensory sampling while animals navigated in darkness, we developed three novel experimental techniques to track movements and detect sensory sampling from a freely behaving WEF: 1) an EOD detector to remotely and accurately measure the sensory sampling rate, 2) an electrical tracking method to track multiple WEF using their own EODs, and 3) visual tracking algorithm for robust body tracking through water under infrared illumination. These techniques were successfully applied to reveal novel sensory sampling behaviours in freely exploring Gymnotus sp. Cortical activity precedes self-initiated movements by several seconds in mammals; this observation has led into inquiries on the nature of volition. Here we demonstrate the sensory sampling enhancement also precedes self-initiated movement by a few seconds in Gymnotus sp. Next, we tested whether these animals can be trained to learn a location of food using electrically detectable landmarks and, if so, whether they can use their past experiences to optimize their sensory sampling. We found that animals revisited the missing food location with high spatial accuracy, and they intensified their sensory sampling near the expected food location by increasing the number of EOD pulses per unit distance travelled.