Research of multi-thread applications for real-time control systems on humanoid robot embedded platforms

摘要

Thanks to the evolution of science technology, lots of applications can be realized in humanoid robots. Race-based humanoid robot's related technologies have also become one of the most important issues in recent years, such as motion control system, environment perception, motion detection, highspeed visual processing, artificial intelligence, knowledge management, and intelligent control system. In order to attain these functions, a control platform is thus designed in this paper to integrate different distributed control systems of race-based humanoid robots. We not only integrate peripheral transmission hardware, but also design different Application Programming Interfaces (APIs) that enable artificial intelligent system to demand the subsystems through our embedded platform and simplify the communication processes among the subsystems. In addition to the above-mentioned systems, a real-time module is further designed to achieve real-time distributed processing. The aim of this paper is to propose a multi-thread programming for processes to cooperate with one another more efficiently. Our experiment results show that this multi-thread platform can decrease the response-time of I/O and enable humanoid robots to react much more rapidly and smoothly.