Sufficiently Accurate Model Learning

Modeling how a robot interacts with the environment around it is an important prerequisite for designing control and planning algorithms. In fact, the performance of controllers and planners is highly dependent on the quality of the model. One popular approach is to learn data driven models in order to compensate for inaccurate physical measurements and to adapt to systems that evolve over time. In this paper, we investigate a method to regularize model learning techniques to provide better error characteristics for traditional control and planning algorithms. This work proposes learning "Sufficiently Accurate" models of dynamics using a primal-dual method that can explicitly enforce constraints on the error in pre-defined parts of the state-space. The result of this method is that the error characteristics of the learned model is more predictable and can be better utilized by planning and control algorithms. The characteristics of Sufficiently Accurate models are analyzed through experiments on a simulated ball paddle system.

Authors

• Clark Zhang
• Arbaaz Khan
• Santiago Paternain
• Alejandro Ribeiro

Keywords

• Task analysis
• Optimization
• Data models
• Adaptation models
• Heuristic algorithms
• Neural networks
• Planning
• Learning Models
• Feature Model
• State Space
• Interior Point Method
• Error Characteristics
• Model Analysis
• Neural Network
• Optimization Problem
• Dynamical
• Artificial Neural Network
• Gradient Descent
• Function Approximation
• Forward Model
• Lipschitz Continuous
• Robotic Arm
• Constrained Optimization Problem
• Pitch Angle
• Roll Angle
• Dual Problem
• Constrained Model
• Duality Gap
• Ball Velocity
• Task Parameters
• Gradient Ascent
• Gradient Descent Step