Krishnamoorthy Kalyanam

Optimization of Perimeter Patrol Operations Using Unmanned Aerial Vehicles

This paper addresseses the following base perimeter patrol problem: a team of unmanned aerial vehicles (UAVs) equipped with cameras and a remotely located operator cooperatively perform the task of perimeter surveillance. There arem alert stations/sites on the perimeter where a nearby breaching of the perimeter by an intruder is flagged byanunattended ground sensor (UGS).Todeterminewhether an incursionflaggedby aUGS is a false alarmora real threat, a patrolling UAV flies to the alert site to investigate the alert. The longer a UAVdwells (loiters) at an alert site, the more information it gathers; however, this also increases the delay in responding to other alerts. The decision problem for aUAV is to determine the optimal dwell time so as tomaximize the expected payoff. In this paper, patrols consisting of one and two UAVs are considered. A stochastic dynamic programming approach is employed to obtain optimal policies for the patrolling UAVs. Theoretical performance bounds from queueing systems literature have been used to benchmark the optimal controller. Also, simulation results for the optimal patrols showing the expected information gained and response time for different alert arrival rates are presented.

Two-Period Repetitive and Adaptive Control for Repeatable and Nonrepeatable Runout Compensation in Disk Drive Track Following

This paper presents the design and implementation of an integrated two-period repetitive and adaptive control scheme to reject both repeatable and nonrepeatable disturbances in the track-following servo control of a hard disk drive read-write head. A baseline linear quadratic Gaussian controller is augmented by a plug-in repetitive controller to reject periodic disturbances with two periods, one that is synchronized with the disk rotation and the other that is not, and an additional adaptive-Q control scheme to reject the remaining aperiodic and random disturbances. The adaptive-Q control algorithm uses the well-known result that all stabilizing controllers for a plant can be synthesized by conveniently parameterized augmentation to a nominal controller. The Q-filter is restricted to a finite impulse response filter, which minimizes the root-mean-square value of the track-following error. Experimental results show substantial performance improvement for the two-stage augmented controller over the single-stage repetitive or adaptive-Q control acting alone.

Experimental Study of Adaptive-

This paper presents the design of an adaptive-Q control algorithm for minimizing the position error signal (PES) in a hard-disk-drive track-following servo. Also, we discuss the results obtained from experiments done on a commercial disk drive using the said scheme. The adaptive- Q control algorithm approach uses the well-known result that all stabilizing controllers for a plant can be synthesized by conveniently parametrized augmentation to a nominal controller. The augmentation to a linear quadratic Gaussian (LQG) optimal control is parametrized by a stable filter Q. The Q filter is restricted to a finite-impulse response filter, which minimizes the rms value of the track following error. Experiments done with a fifth-order adaptive filter that uses PES signal as feedback show a 15% improvement in the achievable track misregistration over a model-based LQG controller. The adaptive-Q algorithm implemented via a recursive least squares numerically stable adaptive algorithm shows fast convergence rate and a consistent performance improvement over a fixed controller for various track locations. Also, the tracking performance improves with increase in the adaptive filter order.

Q

The optimal control of two pursuers searching for a slower moving evader on a Manhattan grid road network is considered. The pursuers do not have on-board capability to detect the evader and rely instead on Unattended Ground Sensors (UGSs) to locate the evader. We assume that all the intersections in the road network have been instrumented with UGSs. When an evader passes by an UGS location, it triggers the UGS and this time-stamped information is stored by the UGS. When a pursuer arrives at an UGS location, the UGS informs the pursuer if and when the evader passed by. When the evader and a pursuer arrive at an UGS location simultaneously, the UGS is triggered and this information is instantly relayed to the pursuer, thereby enabling “capture”. We derive exact values for the minimum time guaranteed capture of the evader on the Manhattan grid and the corresponding pursuit policy.

Control for Disk Drive Track-Following Servo Problem

In this article, a stochastic optimal control problem involving an unmanned aerial vehicle flying patrols around a perimeter is considered. To determine the optimal control policy, one has to solve a Markov decision problem, whose large size renders exact dynamic programming methods intractable. Therefore, a state aggregation based approximate linear programming method is used instead, to construct provably good suboptimal patrol policies. The state space is partitioned and the optimal cost-to-go or value function is restricted to be a constant over each partition. The resulting restricted system of linear inequalities embeds a family of Markov chains of lower dimension, one of which can be used to construct a lower bound on the optimal value function. In general, the construction of a lower bound requires the solution to a combinatorial problem. But the perimeter patrol problem exhibits a special structure that enables tractable linear programming formulation for the lower bound. This is demonstrated and...

Optimal Cooperative Pursuit on a Manhattan Grid

A novel mixed initiative optimal control system for intelligence, surveillance and reconnaissance (ISR) operations which entails human-machine teaming has been developed. The scenario entails a camera-equipped unmanned air vehicle sequentially overflying geolocated objects of interest, which need to be classified as either a true or false target by a human operator. The vehicle is allowed a prespecified number of revisits, such that an object can be looked at, a second time, under better viewing conditions. The overarching goal is to correctly classify the objects and minimize the false alarm (FA) and missed detection (MD) rates. We design a stochastic controller that computes if and when a revisit is necessary and also the optimal revisit state, i.e., viewing altitude and aspect angle. The concept of operation is such that the critical task of detection/pattern recognition is relegated to the human operator, whereas optimal decision making is entrusted to the machine. The stochastic dynamic programming-based decision algorithm is, however, informed about the performance of the human operator via an empirical human perception model. The model is experimentally obtained in the form of state-dependent confusion matrices. The optimal closed-loop ISR system is shown to experimentally achieve a FA rate of 5% and MD rate of 12%, which are significantly lower than the open-loop operator-only performance metrics. The performance improvements that were observed are relevant to a particular operator, and thus, the study suggests that the same improvements could conceivably be achieved with other test subjects.

Lower Bounding Linear Program for the Perimeter Patrol Optimization Problem

We consider the optimal control of a “blind” pursuer searching for an evader moving on a road network with fixed speed toward a set of goal locations. To aid the pursuer and provide feedback information, certain roads in the network have been instrumented with unattended ground sensors (UGSs) that detect the evaders motion. When the pursuer arrives at an instrumented node, the UGS therein informs the pursuer whether and when the evader visited that node. The pursuer is also made aware of the evaders speed. Moreover, the embedded graph comprised of the UGSs as vertices and connecting roads as edges is restricted to being a directed acyclic graph (DAG). The pursuers motion is not restricted to the road network. In addition, the pursuer can choose to wait/loiter for an arbitrary time at any UGS location/node. At time 0, the evaders entry into the road network is registered at UGS 1, the entry node to the graph. The pursuer also arrives at the entry node after some delay

Optimal Human–Machine Teaming for a Sequential Inspection Operation

d

Pursuit of a Moving Target with Known Constant Speed on a Directed Acyclic Graph under Partial Information

and is thus informed about the...

Simulating Cooperative Control Algorithms Using MATLAB, Simulink, and AMASE

with the familiarity of MATLAB. The MATLAB-AMASE Toolbox was created to fulll this need. With a simple interface requiring hardly more than a researcher’s bare algorithm, as well as advancing integration with Simulink, the MATLAB-AMASE Toolbox provides the opportunity to test cooperative control algorithms for UAVs in a realistic environment while saving time and eort on researchers’ behalf. The MATLAB-AMASE Toolbox has already been applied to the Perimeter Patrol Problem successfully and aorded a realistic perspective on algorithm performance with minimal eort.