Improved Data Fusion for Multi-Sensor Tracking using a Reinforced Viterbi Algorithm
Rajarshi Biswas, Akash S. Doshi, Akankshya Bhatta, Sibi Raj B. Pillai · 2019
Employing multiple wide aperture radars with partially overlapping coverage to accurately track moving objects is becoming increasingly popular. However, identifying a common track across the radars can be challenging when each radar sensor obtains multiple measurements from different targets in its field of view. The presence of clutter and spurious measurements further complicates this problem. Data association and target tracking in this context can benefit from the combined processing of the sensor measurements. We adapt the well known single sensor Viterbi Data Association (VDA) algorithm to exchange information between multiple sensors, thereby reinforcing the target tracking performance. The proposed multi-sensor data fusion algorithm is demonstrated to have vastly improved performance over conventional single sensor techniques.