Reduce UAV Coverage Energy Consumption through Actor-Critic Algorithm

摘要

Unmanned aerial vehicles (UAVs) are powerful tools for several applications like transportation and observation. The main reason is the enormous capabilities of such aerial vehicles in terms of mobility, autonomy, communication and processing power at a relatively low-cost. In recent years, due to the continuous development of UAV technology, it has broad prospects in regional coverage application. However, in practical applications, it is very difficult to get an actual mathematical model because of the limited data obtained. Therefore, we chose to use reinforcement learning to solve this problem. In this paper, we form a rule by Reinforcement Learning to cover the coverage of UAVs. We mainly solve two problems: (1) Reduce UAV energy consumption by reducing UAV action times. (2) Solve the huge problem of the dimension space of the value function by using the Actor-Critic algorithm. We compare our method with the traditional coverage method, the result shows that the UAV using the reinforcement learning model consumes less energy often when the same coverage area is completed.