Some of the ideas for the movement and senses of the horsepower robot are coming from the Corpus Animatae pig robot which had main power on the each of the back wheels (or robot legs) - those could move direct forward, backwards and rotate. It's could move randomly in the space, (white) border was a signal for the robot not to cross it.
One of the goals for the horsepower robot was to somehow discuss alternative means of the power as pig robot had a big battery. Limitations of the bioreactor or microbial fuel cells (MFC) powered through decomposing of the horse meat was familiar fact at the beginning - if nothing at all produced energy might act as a switch or it might act as input for the program of the movement of the horse robot.
The construction of the body of the horse was decided in favour to horse animal anatomy. Small robot on the rails or even on wheels just might move on power accumulated through MFC's.
For the horse as quadruped animal (related to all biped and quadruped), the movement depends on the speed - resting on all four legs. In the robotic movement robotized automata with determined path of leg movement without the support could not stand just on two legs (if each part of the leg would be manipulated through servo-motors it just might stand on two, but better on three legs / in theory with inverse kinematics it might stand upright only on one leg if servo-motors would balance its body).
Interest in horse anatomy and its movement were through the process of designing the body related basically to marionette movement (previous posts Warhorse or other mechanical horses that were hanged on strings as mobiles), but in order for horse robotized automata to move the position of the tip of each foot must touch the ground.
marionette design for the leg movement
small scale prototype for analogue determination of the tip of the foot