Siavash Haroun Mahdavi

Awarded VEIV EngD

This work deals with a robot where no domain knowledge is given at all. The genetic algorithms used to control these robots are deeply embodied within their environment. This embodiment means that a very simple genetic algorithm is able to control smart materials like shape memory alloys without any previous domain knowledge about them. The control and exploitation of smart materials is also extended to stereolithographic models. The final outcome of this research will be a robot that is made entirely out of smart materials. This robot will be able to perform a variety of tasks from locomotion, to damage recovery, and even optimising its shape to get the best reception.



  • 2003. Evolving Motion of robot with muscles
  • 2003. An Evolutionary Approach to Damage Recovery of Robot Motion with Muscles.
  • 2003. An Evolutionary approach to microstructure optimisation of stereolithographic models
  • 2003. Evolving Noise Tolerant Antenna Configurations Using Shape Memory Alloys
  • 2004. Blurring the Boundaries between Actuator and Structure: Investigating the use of Stereolithography to build Adaptive Robots.
  • 2004. Optimising Continuous Microstructures: A Comparison of Gradient-Based and Stochastic Methods.
  • 2004. Creativity Through Embodiment: A computational model exemplified by microstructure design
  • 2004. Modularity and Flexibility at the Small Scale: Evolving Continuous Material Variation with Stereolithography.
  • 2004. Innately Adaptive Robotics through Embodied Evolution

Primary Supervisor: Peter Bentley

Industry Sponsor: BAE Systems - Aerospace and Defence

Adaptive robotics