Random Time Delay Effect on Out-of-Sequence Measurements

Abstract

The communication links of multi-agent systems (MASs) generate random time delays, which significantly affect accuracy and real-time control. A Markov chain is established at the input end and output end of a communication link to express this time delay and is introduced to asynchronous data fusion at an intelligent location. The influence of random time delays and data fusion on the positioning accuracy and real-time performance is investigated. A distributed fusion simulation system is established based on MASs and employed to simulate the random delay in the fusion of multi-sensor data. Different fusion algorithms are tested by changing the upper boundary of the time delay. The test results indicate the existence of a boundary between the data fusion and the time delay. The fusion accuracies on both sides of the boundary significantly differ. This difference is helpful for selecting sensors and fusion algorithms for different sampling frequencies in multi-agent collaborative control systems, improving the positioning accuracy of the system and effectively controlling computation costs.