Modern robots are complex heterogeneous systems composed of different Processing Elements (PEs) with multiple sensors and actuators. This implies that different experts are needed to build such systems. Traditionally, robots included Central Processing Units (CPUs) as their PE. However, this has been changing over the last decade as different PEs, namely Graphics Processing Units (GPUs) and Field Programmable Gate Arrays (FPGAs), have drawn the attention of roboticists. The research community focused on various techniques, methodologies, and applications separately, making integration aspects highly complex. Models, as abstractions, have been proposed to aid in designing complex systems that can also help with integration. Hence, three complementary goals are discussed in this work. The first is which robotic applications benefit from parallelizable and energy-efficient devices such as GPUs and FPGAs. The second one is to understand the contributions of different model-based approaches. Lastly, how these two can complement each other to bring benefits from one field onto the other so hardware developers, as well as roboticists, can improve the design of state-of-the-art robotic platforms.