Navigation control of unmanned aerial vehicles in dynamic collaborative indoor environment using probability fuzzy logic approach

Abstract

The development of drones in various applications makes it essential to address the critical issue of providing collision-free and optimal navigation in uncertain environments. The current research work aims to develop, simulate and experiment with the Probability Fuzzy Logic (PFL) controller for route planning and obstacle avoidance for drones in uncertain static and dynamic environments. The PFL system uses probability-based impact assessment and fuzzy logic rules to deal with unknowns and environmental changes. The fuzzy logic system takes in input about the distance of objects from the drone's front, left, and right sides, as well as the probability of collision based on the drone's speed and how close it is to the obstacles. The set of thirty fuzzy rules based on the distance of the obstacle from front left and right are defined to decide the output, i.e. speed of the drone and heading angle. The simulation environment is developed using MATLAB, with grid-based motion planning that accounts for static and dynamic obstacles. The system's performance is validated through simulations and real-world experiments, comparing path length and travel time. On comparing the simulation and experimental results, the proposed PFL-based controller has been proven to be efficient, accurate, and robust for both static and dynamic and simple to complex environments. The drones can plan the shortest and most collision-free path across all the scenarios, as depicted in the simulation and experimentation results. However, due to communication delay, inaccuracy of sensor response, environmental impact and motor delay, there are slight deviations between the simulation and experimentation values. Upon performing the error analysis, it is found that the error between the simulation and experimental value is within the range of 6.66 % in all the studied scenarios.