Selection and scheduling of perception resources based on group intelligence perception technology in network autopilot

In order to ensure that the representative car can guarantee the quality of the perceived information while minimizing the total cost of the perceived task in the group intelligence perception information of common perceptual road conditions. In view of the above problems, this paper proposes a combination of greedy algorithm and genetic algorithm to compare the vehicle selection method generally used to solve the knapsack problem. There are many improvements in the overall data reliability of the representative car. The experimental results show that the overall data reliability of representative vehicles has been greatly improved, the total task processing time is reduced and the average perception data reliability of representative vehicles is improved.