Mixed Reality Arm Motion Synthesis Technology Based On Keyframe And Inverse Kinematics

In recent years,Mixed Reality(MR)technology has made great progress in both basic hardware and application.In mixed reality under the original sensor devices arm motion synthesis,mixed reality devices because only low dimensional human hand motion information collected,so most of the mixed reality application only shows the hand movement,without the movement of the elbow and shoulder joints,said for emphasizing the immersion and authenticity of mixed reality technology is obviously not enough.At present,to generate high-quality arm movement under mixed reality equipment,most schemes use additional sensors to obtain higher dimensional movement data,but it also increases the difficulty of setting up the environment,so that users wear more devices.Therefore,it is necessary to use the low-dimensional features of original sensors of mixed reality devices to synthesize continuous and natural arms at the expense of a small amount of arm motion generation accuracy.When the human body moves in some specific scenes,the motion track of each joint is fixed and the motion process is repeated,such as dance,martial arts and other sports.Therefore,a scheme of arm motion synthesis based on the hand position information captured by mixed reality equipment is proposed.Firstly,it is advantageous for the nearest neighbor propagation clustering algorithm based on the physical structure features of human body to automatically extract the motion key frame,so as to reduce the large amount of redundant data in human motion data.Secondly,the method only needs to provide the wrist position,which is regarded as the end effector position.The arm position and attitude are approximated by the joint rotation Angle solved by inverse kinematics,so as to synthesize high-quality human arm motion sequence.The simulation results show that this method shorts the backtracking path of FABRIK inverse kinematics algorithm and reduces the amount of computation.The proposed method drives the motion fast,and improves the inverse kinematics algorithm’s approximation to the end effector but cannot generate highquality arm motion.It solves the problem that it is difficult to reconstruct natural and continuous virtual arm with real-time data under a few sensors in mixed reality application,and can be used to synthesize high-quality virtual arm movement in mixed reality application.However,when the type of hand movement is not in the movement database,only the inverse kinematics algorithm will be used to synthesize the arm movement,but there are still some problems such as unnatural synthesis movement and large amount of calculation.