A genetic programming framework for error recovery in robotic assembly systems

Abstract

In this paper, the advantages and performance of genetic programming in use of error recovery planning in robotic assembly systems is discussed. Existing systems use polynomial time planning techniques or heuristics to produce error recovery plans. However, these systems require translation of the generated plans to working controller codes. An alternative approach could be the use of Genetic Programming to produce recovery plans in robot language itself. A framework is developed and coupled with a 3D robotic simulation software for the generation of error recovery logic in assembly systems. The developed architecture uses a genetic programming system based on both deterministic and probabilistic crossover and variable mutation schemes. Performance of the system is evaluated with the simulations made on a three dimensionally modeled automated assembly line. The obtained results showed that the deterministic crossover operator improves the evolution of the plans and the system is efficient of generating robust recovery plans for different error states.