In this paper a minimal control strategy for the reproduction of steering strategies in insects is reported. The strategy is a further development of a control methodology for locomotion control based on the Central Pattern Generator and accompanied with rigorous stability analyses obtained exploiting the Partial Contraction Analysis method. The mathematical model was derived extending the CPG network already introduced by the authors with an additional degree of freedom in each leg of a simulated legged robot, augmenting the leg steering possibilities for better mimicking Neurobiological findings. The control methodology aims at implementing the steering strategy using, at the same time, a minimal number of control variables, which are related to the phase displacements among the oscillators of the femur and tibia of the robot legs. Experimental results carried out in a dynamic simulation environment are also reported.