We study the problem of sensor scheduling for multisensor multitarget tracking-to determine which sensors to activate over time to trade off tracking error with sensor usage costs. Formulating this problem as a Partially Observable Markov Decision Process (POMDP) gives rise to a non-myopic sensor-scheduling scheme. Our method combines sequential multisensor Joint Probabilistic Data Association (MS-JPDA) and particle filtering for belief-state estimation, and uses simulationbased Q-value approximation method for "lookahead." The example of focus in this paper involves the activation of multiple sensors simultaneously for tracking multiple targets, illustrating the effectiveness of our approach.
The forward kinematics of manipulators are frequently represented as to potentially catastrophic incidents in remote and/or hazardous environments. A direct approach towards increasing robot reliability is to improve the reliability of the individual components; however, achieving acceptable reliability rates is often prohibitively expensive, and sometimes technologically impossible. An alternative approach is to consider failure-tolerant robot designs. These typically incorporate a failure detection and identification scheme [:3J followed by failure recovery [4J. Designing the robot with redundant systems increases the options available for failure tolerance. Redundancy can be in the form of duplication of actuators and sensors [5, 6J, or in the form of kinematic redundancy [7, 8, 9J. Proper utilization of kinematic redundancy provides greater dexterity prior to failures, minimizes the immediate impact of a failure, and guarantees task completion by ensuring a reachable postfailure workspace.This work presents real-time implementation of schemes for utilizing kinematic redundancy to maximize the tolerance of robots to single locked-joint failures. The fault-tolerance measure used is a worst-case quantity, given by the minimum, over all single joint failures, of the minimum singular value of the post-failure Jacobians. Maximizing this measure corresponds to configuring the robot to minimize the worst-case end-effector velocity error over all single locked-joint failures. This is also equivalent to minimizing the worst-case discontinuities in the joint velocities of the healthy joints as they move to compensate for the failed joint. The main contributions of this work are computationally efficient schemes for computing the fault-tolerance measure and its gradient, which in turn enable real-time optimal configuration of robots in anticipation of failures.where x is an m-dimensional vector representing the end-effector velocity, q is an n-dimensional vector describing the joint velocities, and J is the m by n manipulator Jacobian matrix [10]. For a redundant manipulator, n > m so the equation is underconstrained and there are an infinite number of solutions which can be expressed as ABSTRACT This work considers real-time fault-tolerant control of kinematically redundant manipulators to single locked-joint failures. The fault-tolerance measure used is a worst-case quantity, given by the minimum, over all single joint failures, of the minimum singular value of the post-failure Jacobians. Given any endeffector trajectory, the goal is to continuously follow this trajectory with the manipulator in configurations that maximize the fault-tolerance measure. The computation required to track these optimal configurations with brute-force methods is prohibitive for real-time implementation. We address this issue by presenting algorithms that quickly compute estimates of the worst-case fault-tolerance measure and its gradient. Real-time implementations are presented for all these techniques, and comparisons show that the perfo...
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.