A biped gait learning algorithm for humanoid robots based on environmental impact assessed artificial bee colony

Gait pattern performance is a very important issue in the field of humanoid robots, and more and more researchers are now engaged in such studies. However, the tuning processes of the parameters or postures are very tedious and time-consuming. In order to solve this problem, an artificial bee colony (ABC) learning algorithm for a central pattern generator (CPG) gait produce method is proposed in this paper. Furthermore, the fitness of the bee colony is considered through environmental impact assessment, and it is also estimated from the cause of colony collapse disorder from the results of recent investigations in areas, such as pesticides, electromagnetic waves, viruses, and the timing confusion of the bee colony caused by climate change. Each environmental disaster can be considered by its adjustable weighting values. In addition, the developed biped gait learning method is called the ABC-CPG algorithm, and it was verified in a self-developed high-integration simulator. The strategy systems, motion control system, and gait learning system of the humanoid robot are also integrated through the proposed 3-D simulator. Finally, the experimental results show that the proposed environmental-impact-assessed ABC-CPG gait learning algorithm is feasible and can also successfully achieve the best gait pattern in the humanoid robot.