Unmanned Systems Technology 024 | Wingcopter 178 l 5G focus l UUVs insight l CES report l Stromkind KAT l Intelligent Energy fuel cell l Earthsense TerraSentia l Connectors focus l Advanced Engineering report

54 Report | Consumer Electronics Show 2019 AutoX came to the show with a fleet of its self-driving grocery delivery cars for viewing and to provide autonomous fast-food deliveries throughout the event. The company’s autonomous navigation and perception uses sensor fusion of a 360° Lidar, six radar modules, eight vision cameras and GPS, as well as persistent data links to the AutoX fleet command centre. A total of six antennas are mounted atop each vehicle to receive GPS signals and provide 4G LTE comms (and redundancy for both). A ‘black box’ drive is kept on board each vehicle to keep recordings of the full stack of data. This includes the vehicle’s position, its velocity, Lidar data, camera imagery, radar information and corresponding outputs such as perception, vehicle decisions, planning and control data. These can span the length of a test drive or just for three minutes before and after being manually triggered by the onboard safety driver, or automatically triggered by the autopilot. This data (including those in which crashes or collisions occur) can later be fed into AutoX’s simulation programs in order to more closely identify and isolate potential points for improvement in the sensors or AI, in the company’s pursuit of SAE Level 4 autonomy. Intrepid Control Systems unveiled its Autonomous ONE motherboard, designed for powering self-driving car intelligence. The system is intended to provide a single interface drive to accelerate driverless vehicle progress, compared with typical development roadmaps that rely on integrating a wide range of systems from different companies to interface with each other. Each ONE board will include a CPU card for autonomy that supports PCIe, MIPI camera inputs, Ethernet and USB (with the company having tested the new system with Nvidia Jetson TX2 and Xavier modules). “The board is designed to fit any standardised 1U 19 in rack,” said Arpan Rughani. Four USB 3.0 ports are installed to support UART or camera sensor inputs. Several other I/O ports are also provided, for up to four additional cameras, radar, Lidar and other sensors. To allow developers to analyse and improve their vehicles, four NVMe M2 (or M3) cards can provide up to 64 Tbytes of removable data storage. Positioning updates are given by an integrated Ublox Neo-M8 receiver, and time synchronisation of sensor and positioning data can be programmed using CANFD, Ethernet (PTP/NTP/ Intrepid) or GPS PPS. French start-up Notilo Plus showcased the iBubble, an AUV developed for underwater videography and photography to support the work of divers. It measures 50 x 35 x 25 cm and weighs about 9 kg with batteries. February/March 2019 | Unmanned Systems Technology Food on the go, with self-driving grocery delivery cars from AutoX The Autonomous ONE motherboard is for powering driverless car intelligence