Dynamic Partial Reconfigurable FPGAs (DPRFPGA) are integrated programmable circuits that are dynamically reconfigurable: the designer can reconfigure them runtime, depending on the (dynamic) environment. This paper presents a new model-based design flow for DPR-FPGAs, based on the creation of a specific Extended Finite State Machine (EFSM) which formally describes the hardware. The design flow uses Pyngu, a tool which automatizes the generation of code and simulation through SystemC. This methodology is applied to a reconfigurable robot deployed in different scenarios, which can reconfigure its kinematics in order to overcome obstacles: the kinematics depends on the robots shape, enabling different movement types like trotting, crawling and getting up. Experiments show that the model-based design flow leads to effective adaptable robots design.
Titolo: | A model-based design flow for Dynamic Partial Reconfigurable FPGAs |
Autori: | PRAVADELLI, Graziano (Corresponding) |
Data di pubblicazione: | 2019 |
Handle: | http://hdl.handle.net/11562/1001100 |
Appare nelle tipologie: | 04.01 Contributo in atti di convegno |