In summary both RR and RP robots can exhibit chaotic motions. Chaotic motions typically were found to be for low controller gains, chaotic motions for large mismatch between model and plant, chaotic motions were also seen for if the model was underestimated and most of the time we could see that the route to chaos was through period doubling and one more conclusion, in spite of all our simulations, we could never see chaos in a robot which consist of a P and an R and a P and a P combination. Finally, I think it is important for us to study chaotic motions because it goes against the basic fundamental results that all controllers, a PD controller or a model-based controller, will ensure that the trajectory or the output is asymptotically approaching the input, whereas this simulation clearly shows that although the output is bounded it need not be approaching the input. So, after a while your output is unpredictable.

---