Novel Sensor Scheduling Scheme for Intruder Tracking in Energy Efficient Sensor Networks

Distributed Sensor Management Based on Target Losing Probability for Maneuvering Multi-Target Tracking

“…The comparison results of method 2, method 3 and method 4 are shown in Fig.13. Fig.13(a) shows the changes of objective function values in equation (13); Fig.13 (b) shows the changes of radar emission interception probability; Fig.13 (c) shows the changes of target tracking accuracy; Fig.13 (d) shows the changes of target losing probability.…”

“…It can been seen from Fig.13(a) that the objective function values in method 4 keep almost the lowest among the three sensor management methods at most time instants, the sensor management method in the constraint of target losing probability (method 1) performs worse than the sensor management based on information gain (method 4) at most time instants, and the randomly controlling based method (method 3) behaves the worst at most time instants. The reason for this result is that for the objective function, both the target tracking accuracy and radar emission interception probability are added in method 1 as equation (13) shows. However, in method 4, only the target tracking accuracy is taken into consideration and optimized to the greatest extent.…”

“…It can been seen from Fig.13(b) that the radar emission interception probability values in method 1 keeps almost the lowest at most time instants, method 3 performs worse than method 1, and method 4 behaves the worst among three sensor management methods. The reason for this result is that only in method 1, the radar emission interception probability is optimized shown as equation (13). And in method 4, more radars are used to improve the target tracking accuracy, leading to the biggest radar emission interception probability.…”

“…For the both two categories of sensor management, the method of building the objective function in sensor management can be concluded as follows [13], [14].…”

A Markov-Based Intrusion Tolerance Finite Automaton

“…Diddigi et al [18] put forward a reinforcement learning algorithm that tracks intrusions to the tolerance system through upper confidence tree search. With the aid of Markov decision process, the algorithm optimizes the state space and the action space by accurately calculating sensor data, and pinpoints the intrusion nodes in the tolerance system, thereby maintaining the network security at a high speed.…”