Research on Improved Multi-Sensor Data Fusion Algorithm Based on D-S Evidence Theory
Junsuo Qu, Xing Cai, Haonan Shi · 2021 4th International Conference on Artificial Intelligence and Pattern Recognition · 2021
The data fusion of multi-sensor enhances the intrinsic relationship of data between sensors, reduces the workload of information processing, and will not miss the important information features. This paper points out the main problems in the application of D-S evidence theory, analyzes and compares the existing improved methods. When the evidence conflict is large, the traditional D-S evidence theory synthesis formula is inconsistent with the actual situation and the result is invalid. The priority factor is introduced to reallocate the basic probability assignment of the evidence conflict part. And the method is used for map construction simulation. From the constructed ring map, it can be seen that the obstacle points of each decision are very close to the obstacle points in the simulation. The reliability and accuracy of the method of introducing priority factor in drawing construction are explained.