Evolutionary Optimization of Mechanical and Control Design Application to Active Endoscopes

Simultaneous structure/control optimization in a robotic system design is addressed through Evolutionary Algorithms (including genetic algorithms and genetic programming). Both aspects are evolved in the same evolutionary algorithm through dynamic simulations and simulation approximations for continuous and task oriented evaluations. Here, we investigate a specific adaptation of these principles to the design of smart active endoscopes for minimally invasive diagnosis or surgery. The design of such mini-robotic systems is based on a serial arrangement of articulated rings with associated antagonist SMA micro-actuators which configuration have to be adapted to the surgical operation constraints. The control strategies for adaptation of the system geometry to the environmental constraints are based on a multi-agent approach to minimize the inter-module communication requirements. The results obtained from the developed Genetic Algorithm-based design software for the particular application of colonoscopy show the consistency of the solutions. Moreover, the proposed technique for synthesis of approximated evaluation functions significantly speedup the design process while leading to robust fitness representations.