Data-Driven Multi-step Nonlinear Model Predictive Control for Industrial Heavy Load Hydraulic Robot

TL;DR

Simulations and experiments on a 22-ton hydraulic excavator have validated the effectiveness of the proposed NMPC approach, showing that the proposed NMPC approach can be widely applied to industrial systems, including nonlinear control and energy management.

Abstract

Automating complex industrial robots requires precise nonlinear control and efficient energy management. This paper introduces a data-driven nonlinear model predictive control (NMPC) framework to optimize control under multiple objectives. To enhance the prediction accuracy of the dynamic model, we design a single-shot multi-step prediction (SSMP) model based on long short-term memory (LSTM) and multilayer perceptrons (MLP), which can directly obtain the predictive horizon without iterative repetition and reduce computational pressure. Moreover, we combine offline and online models to address disturbances stemming from environmental interactions, similar to the superposition of the robot's free and forced responses. The online model learns the system's variations from the prediction mismatches of the offline model and updates its weights in real time. The proposed hybrid predictive model simplifies the relationship between inputs and outputs into matrix multiplication, which can quickly obtain the derivative. Therefore, the solution for the control signal sequence employs a gradient descent method with an adaptive learning rate, allowing the NMPC cost function to be formulated as a convex function incorporating critical states. The learning rate is dynamically adjusted based on state errors to counteract the inherent prediction inaccuracies of neural networks. The controller outputs the average value of the control signal sequence instead of the first value. Simulations and experiments on a 22-ton hydraulic excavator have validated the effectiveness of our method, showing that the proposed NMPC approach can be widely applied to industrial systems, including nonlinear control and energy management.

Figures (10)

• Figure 1: Power source, joint and actuation system of hydraulic excavator. The final joint requires selecting the appropriate tool hand based on the task. To adapt to grasping tasks, we change to a clamping jaw.
• Figure 2: Offline SSMP model. The LSTM includes input gates, forget gates, and output gates, with its input being the system's state. The input for MLP is the system's control signals, which must be transformed into a one-dimensional format if they are originally multidimensional.
• Figure 3: Data Collection Method. The workspace is a subset of the robot's reachable space. The safety monitor calculates the joint cartesian coordinates and determines whether it exceeds the workspace. The pre-configured controller is typically the controller provided by the manufacturer.
• Figure 4: Block diagram representation of the NMPC. After the manager issues a task, trajectory planning and online model resetting are initiated concurrently. The red block contains the main parts of NMPC, including the offline model, online model, online learning, and online optimization. Solid arrows represent data exchange between different components, while hollow arrows indicate data transfer within NMPC.
• Figure 5: The experimental machine and corresponding simulation model. The simulation model is constructed based on AMESim demo and using the same sensors as the actual robot. We proactively introduce nonlinearities such as dead zones, leakages, and friction, and optimize the control signals in real-time within Simulink s-function.