Evolving Vascular Morphogenesis Controller to Demonstrate Locomotion

Payam Zahadat, Thomas Schmickl
Artificial Intelligence and Signal Processing (AISP), Shiraz (in press) (2017)


Locomotion can be a result of coordination of movements between well-shaped limbs of an organism, or it can be an emerging effect of local interactions between several simple agents forming an aggregated swarm that can move around in the environment. In this paper, we apply a distributed morphogenesis approach to develop a mobile growing artificial organism. The morphogenesis approach works based on a competition between several agents at growing positions that strive to get larger fractions of a limited shared resource needed for growth. The morphology develops from the local competitions and interactions of such agents with each other and with their environment. The locomotion driven by this approach is the result of growing new agents at one side and retracting agents from the other side of the overall swarm-based organism. The morphogenesis controller is successfully evolved here for generating a growing organism that

moves towards a target.

Download PDF: