Iterative Jerk Motion Engine for SCARA Arms
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing motion control systems in closed-loop systems, such as those used in 3D printing, face challenges with high computational requirements, accuracy, and speed, often resulting in slow feedback due to stability issues and the need for precise measurements, which can lead to errors and increased machine size and cost. Additionally, prior art methods struggle with handling non-orthogonal axes and time-varying inputs, and do not effectively utilize system velocity and acceleration for feedback.
Innovation Solution
The implementation of an iterative jerk method that computes a numerical 'jerk' using real-time feedback from the mechanical system, allowing for improved motion engine outputs and motor control by considering current and target positions and velocities, and optimizing for smooth acceleration and velocity profiles without relying on prior motion data or memory.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If PID controllers with feedback from tool head or work piece are used, then closed loop control is achieved, but the system becomes slow due to stability requirements and long delay between motion commands and actual position
Solution Approach 1:
The patent replaces traditional mechanical feedback systems (sensors at tool head) with a computational model-based feedback system. The motion engine uses a dynamic model of the mechanical system to predict actual position and velocity, eliminating the need for physical sensors and long feedback loops while maintaining closed-loop control stability.
Solution Approach 2:
The patent introduces a computational model as an intermediary between the motion commands and actual system state. This model predicts system behavior and provides intermediate feedback signals that guide control decisions, avoiding direct reliance on slow physical feedback from sensors.
2Measurement precision
If very-high precision measurement directly at the tool head is implemented, then measurement accuracy is improved, but machine size and cost increase due to rigidity requirements
Solution Approach 1:
The patent replaces physical measurement devices (encoders, lasers, cameras) with a computational sensing system. The motion engine's dynamic model mathematically determines tool head position and velocity from motor commands and measured motor positions, eliminating expensive precision measurement hardware while maintaining high accuracy.
Solution Approach 2:
The patent creates a virtual copy of the mechanical system in the form of a computational model. This digital twin replicates the physical system's dynamics and provides all necessary measurement information without requiring physical sensors, reducing hardware complexity and cost.
3Extent of automation
If computational sequence from move instructions through motion engine is used, then motion control is achieved, but computational requirements are high requiring both high computational accuracy and high-speed processing
Solution Approach 1:
The patent performs preliminary computation of the dynamic model parameters and prediction equations during system initialization or offline setup. This pre-computation reduces the real-time computational burden during motion execution, allowing high-accuracy control with reduced processing power requirements.
Solution Approach 2:
The patent uses a dynamic computational model that adapts to changing system conditions in real-time. The model updates its predictions based on current motor positions and commands, providing accurate feedback without requiring excessive computational resources through efficient algorithms.
4Reliability
If I (integration) term in PID controller is used, then memory of prior operation is maintained, but large errors or slow settling time occur for time-varying inputs
Solution Approach 1:
The patent inverts the traditional feedback approach by using a forward-looking dynamic model to predict future system state based on current inputs, rather than reacting to past errors. This model-based prediction eliminates the need for integrative memory terms that cause slow response to time-varying inputs, while maintaining stability through accurate system modeling.
Data Source
AI summary
Described is a method and device of a computational motion engine iteratively computing a numerical “jerk,” the motion derivative of acceleration, using real-time feedback from a system under motion control, to reach both a desired position and desired velocity of a next waypoint. Output from the motion engine is only desired acceleration, which is then passed to a motor driver, free of intermediate computations of either position or velocity. A second, inside feedback loop maintains desired acceleration or torque at the motor shaft based on the acceleration output of the motion engine, which may use non-linear correction tables. Waypoints comprising both position and velocity are inputs to the motion engine. Time to next waypoint is computed rather than provided as an input. Optimization of moves to the next waypoint is based on smoothest velocity change during the move. Embodiments include mechanical, two-axis SCARA arm motion systems.