Legged Robot Jump Control Using Run-Up Velocity and Action Library
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for controlling the jump of legged robots are limited by the rotation speed of robot joints and force torque, making it difficult to achieve ideal jump heights and are not applicable to legged robots, as they typically control takeoff from a stationary state.
Innovation Solution
A motion control method for legged robots that involves obtaining a jump parameter including expected velocity and height, retrieving a corresponding jump trajectory from a preset action library, and controlling the robot to reach the expected velocity before jumping, using a cost function model constrained by kinetic, velocity, or mechanical models to optimize the jump action.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If the robot controls jump from a stationary state, then the control method is simple, but the jump height is limited by joint rotation speed and force torque
Solution Approach 1:
The robot performs a run-up phase before jumping to accumulate kinetic energy. The control system pre-calculates the required velocity and trajectory, then guides the robot to reach that state before executing the jump, rather than attempting to generate all jump energy from a stationary position.
Solution Approach 2:
The control method transitions from controlling only position to controlling both position and velocity. By managing the velocity state during the run-up phase, the system can achieve higher jump heights while maintaining manageable joint rotation speeds during the actual jump execution.
2Speed
If the robot increases jump height by increasing joint rotation speed, then the jump effect improves, but the system becomes limited by mechanical constraints
Solution Approach 1:
The robot accumulates kinetic energy through a run-up phase before the jump, reducing the need for extremely high joint rotation speeds during the jump itself. This preliminary acceleration allows the system to achieve higher jump heights while respecting mechanical speed constraints.
Solution Approach 2:
The control system dynamically adjusts the robot's velocity and trajectory during the run-up and jump phases. By managing the dynamic state transitions, the system can adapt to mechanical constraints while optimizing jump performance for different height requirements.
3Manufacturing precision
If the robot uses real-time trajectory calculation for jump control, then the jump precision improves, but the computational pressure increases
Solution Approach 1:
The system pre-calculates jump trajectories offline and stores them in an action library organized by velocity and height parameters. During real-time operation, the controller only needs to retrieve the appropriate pre-computed trajectory based on current conditions, dramatically reducing computational pressure while maintaining precision.
Solution Approach 2:
The control approach segments the jump control into discrete states (run-up, takeoff, flight, landing) with specific velocity and height ranges. Each state has pre-computed trajectories stored in the action library, allowing efficient lookup and execution without real-time optimization calculations.
4Adaptability or versatility
If the robot establishes an action library with multiple jump trajectories, then the jump adaptability improves, but the memory requirement increases
Solution Approach 1:
The action library organizes trajectories by key parameters (velocity, height, takeoff angle) rather than storing all possible trajectories. This parameter-based organization allows efficient storage and retrieval, providing high adaptability while minimizing memory requirements through systematic data structuring.
Data Source
Figure 1~2
Figure 3
Figure 4
AI summary
The present disclosure provides a motion control method and apparatus, a robot and a non-transitory storage medium. The method includes: obtaining a jump parameter of the robot, the jump parameter including an expected velocity and an expected jump height before take-off of the robot; obtaining a jump trajectory corresponding to the jump parameter in a preset action library, the action library including a jump trajectory that is marked with a corresponding jump parameter; and controlling the robot to reach the expected velocity, and controlling the robot to jump according to the jump trajectory.