Abstract: Data provided by sensors is always affected by some level of uncertainty or lack of certainty in the measurements.Combining data from multiple sources using multi-sensor data fusion algorithms eliminates the false or incorrect data and exploits the data redundancy to reduce this inconsistency from sensors.Proposes an novel approach to multi-sensor data fusion that relies on combining a Improved Bayesian Approach (IBA) with Particle Filter Algorithm (PFA).Three different methods namely:Pre-Particle Filtering (P-IBA),Post-Particle Filtering (IBA-P) and Pre-Post-Particle Filtering (P-IBA-P) are described based on how filtering is applied to the sensor data,to the fused data or both.A case study of multi-sensor data fusion using optical algorithms is presented.Experimental shows that the shortest running time t was the P-IBA of the decentralized system,overall,the P-IBA-P had the least value of the evaluation function EF and then follows it the IBA-P with a very small difference.This is a reasonable result because the presence of the particle filters,produces estimates with less noise than the measurements and close to the accuracy of the theoretical states.combining Bayesian fusion with particle filtering helps in handling the problem of inconsistency of the data in both centralized and decentralized data fusion architectures.it would be recommended to use P-IBA-P in applications where time is an important factor.

Key words: multi-sensor data fusion, Improved Bayesian Approach (IBA), Particle Filter Algorithm (PFA)