Control algorithm of industrial robot

The algorithm of industrial robot is divided into perceptual algorithm and control algorithm, further subdivided into environment-aware algorithm, path planning and Behavior decision algorithm (AI), motion control algorithm, and the latter two can be collectively referred to as control algorithm

Environment-aware algorithm obtains the environment various data (robot vision and image recognition), locates the robot azimuth (slam), for the fixed station robot, the environment perception algorithm is often not necessary, but for another kind of robot, for example the sweeping robot, is basically a slam algorithm, Behavior decision and motion control algorithm and its simplicity can be neglected
Robot's own motion control algorithm is the focus of research and development of robot manufacturers, mainly to improve the accuracy of the robot operation, stability and speed, the half by PID Servo motor, half by the control algorithm, the same performance of the PID servo motor, a good control algorithm can improve the accuracy of more than 10 times times, the hardware is not the difficulty, Because robot manufacturers all over the world buy the same chip and hardware circuitry.

In general, the environment-aware algorithm and motion control algorithm is more mature, but also the whole robot research field to put the most manpower, and constantly improve the existing algorithm optimization, one is because the research has been a breakthrough, follow the team more, two is because 90% of the robot application field, Only one of the two algorithms is needed, and the behavioral decision algorithm is simple, repeating one or a few simple actions

Behavioral decision-making algorithm or behavior control strategy is not a breakthrough in the field of robotics research and development focus (not the application domain algorithm is different, of course, can also be completely manual control, we often say that the artificial intelligence, narrow point refers to this module), here is not the simple behavior algorithm, such as repetitive action, Robots to dance in fixed motion, barrier-free or fixed-line walking, etc., these are mainly hard-coded implementations, not involving AI, complex behavioral decision-making algorithms are mainly FSM, AHP, decision Tree, Fuzzy Logic, Genetic algorithm GA, Ann, Artificial neural network, and specific problems of the particular algorithm, such as Path planning (Ros provides a move-base module, implemented a number of path planning algorithms), generally used in C + + mixed Python programming

Behavioral decision algorithm, there are good solutions, such as the navigation path planning algorithm, but also very difficult, such as the obstacle avoidance algorithm, almost all the unmanned and autonomous driving research and development team are struggling to think about the obstacle avoidance algorithm, in fact, the application of the obstacle avoidance algorithm is widely used, many areas such as unmanned aerial vehicles to use, The obstacle avoidance algorithm is a stumbling block for the entire driverless and autonomous driving industry, as it determines the last 1% security, while the existing VFH obstacle avoidance algorithm and the DWA obstacle avoidance algorithm can only be very primitive, completely unable to meet the actual needs of the

  No matter the motion control algorithm, Or behavioral decision-making algorithms, German companies are standing at the top of the industry, such as Siemens, Schneider