Result on Game 1. Evolving to max sum of Device/SubDeviceList/InertialSensor/AccelerometerZ/Sensor/Value

Game 1: Selects for higher sum over Z accelerometer over the 10 cycle trial. 

Behavioural Phenotype: This should be maximised by standing on its head. 

Device/SubDeviceList/InertialSensor/AccelerometerZ/Sensor/Value is the Z axis accelerometer. We evolve Nao on the game which is the sum of this sensory state over 10 iterations of the actor molecule loop.

Pop Size = 30. Initial molecules are sensory to motorIzh to motorIzh with random sensory input (length 2) and random motor output (length 3). The perceptron weights are initially random from -0.5 to 0.5.

Fitness over 1000 evaluations (2 hours of behaviour in real time)

Action molecule, start = light blue sensory atom, motorIzh atoms in blue. There are delays between execution of motor atoms.

Some behaviour at the end of the run. There is no great complexity of motor action by this stage in evolution.