Abstract: The motion law of astronauts in spacesuit is studied, and the process of carrying objects is classified. Combined with the traditional six stage acceleration and deceleration algorithm, the muscle activation in each case is planned and solved by the particle swarm optimization algorithm. Using the Matlab/Simulink to build the upper limb skeletal muscle model, the velocity planning of the target object is realized. Changing the dynamic acceleration constraint into the static muscle activation constraint, a velocity planning algorithm based on the muscle activation degree under the dynamic constraints is proposed. The proposed algorithm is verified on the Matlab to prove the feasibility. Compared with other algorithms, the planned muscle activation is more consistent with human biomechanics, which improves the training efficiency of the astronauts.

Key words: Velocity planning, Astronaut virtual training, Dynamic constraint, Muscle activation planning