ing biological development. Alternatively, engineers can draw on insights from biological development to build better robots.

Evo-devo-robo. Developmental robotics tends to focus on post-natal change to a robot’s “body” and “brain” as the robot learns to master a particular skill. Evolutionary robotics experiments on the other hand generate robots that become more complex from generation to generation, but typically each individual robot maintains a fixed form while it behaves.

Biological systems however exhibit change over multiple time scales: individual organisms grow from infants into adults, and the developmental program that guides this change is in turn altered over evolutionary time. This process is known as the evolution of development, or evo-devo. This biological phenomenon has recently been exploited in evolutionary robotics: 3 At the outset of evolution, robots change from a crawling worm into a legged walking machine over their lifetime. As evolution proceeds, this infant form is gradually lost until, at the end of evolution, legged robots exhibit the ability to walk successfully without the need to crawl first. It was found this approach could evolve walking machines faster than a similar approach that does not lead robots through a crawling stage.

In the initial experiments of evo-devo-robo, 34 the genetic instructions were encoded as a specific class of formal grammars known as Linden-mayer systems, or L-systems.c L-systems were initially devised to model plant growth: their recursive nature can produce fractal or otherwise symmetric forms. Hornby12 demonstrated that robots evolved using such grammars do indeed produce repeated forms (Figure 1a). He also showed this repetition can make it easier for evolutionary algorithms to improve such robots, compared to robots lacking in genetically determined self-similarity.

The evolution of robot bodies and brains differs markedly from all other approaches to robotics in that it does not presuppose the existence c Sims’ work had a large impact on the computer graphics community and L-systems remain a popular technique within that field.

the evolution
of robot bodies
and brains differs
markedly from
all other
approaches
to robotics in
that it does not
presuppose
the existence of
a physical robot.

of a physical robot. Rather, the user provides as input a metric for measuring robot performance along with a simulation of the robot’s task environment, and the algorithm produces as output the body plan and control policy for a robot capable of performing the task. This can then be used to manufacture a physical version of the evolved robot. Such an algorithm could, in principle, continually receive new desired behaviors and task environments and continuously generate novel robots.

In this way, the roboticist can make fewer assumptions about the final form of the robot and have greater confidence the final evolved robot is better adapted to the environment in which it must operate. For example, there is often a debate about whether a wheeled or legged robot is more appropriate for moving over a given surface. Although not yet demonstrated, an evolutionary robotics algorithm should generate wheeled robots if supplied with a simulation of flat terrain and legged robots if supplied with a simulation of rugged terrain. Recent work in mainstream robotics has demonstrated the possible advantage of combining wheels and legs in the same robot: an evolutionary system should rediscover this manually devised solution if it is indeed superior to either wheels or legs alone.

Another advantage of this approach over mainstream robotics is its potential for better scalability: by genetically encoding assembly instructions rather than the blueprint of a robot, more complex machines can be evolved with little or no increase in the amount of information encoded in the genome. For example, consider an approach in which robots are specified by a formal grammar such that the invocation of a rewrite rule replaces one part of the robot with two or more parts. Thus the more times a given set of rewrite rules are invoked, the more complex the resulting robot becomes. If evolution increases the number of rewrite rule invocations, then simple robots can evolve into more complex robots with no increase in the information content of the underlying genomes describing those robots.

Despite the promise of this ap-