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.

Augmented reality (AR) applications greatly benefit from direction estimation. In current AR devices, direction estimation is enabled by power hungry technologies such as GPS, management of inertial data and video processing. We present a low-complexity eyewear prototype capable of obtaining directional information from a set of environmental radio sources by using a multi-antenna system patented by Wagoo LLC. Eyewear devices are the natural match for this technology under the assumption that user line of sight represents the main reference for user interaction and orientation. In the demonstration, the prototype detects when a user aims at objects or people that are associated to standard Bluetooth Low Energy emitters (beacons) with a current 30° sensitivity and 5% error rate.