Accurate positioning achieved for autonomous vehicles
30 JANUARY 2020
Autonomous vehicles and advanced driver assistance systems need robust and precise positioning to enable reliable operations. This is especially important in the early transitional phase of the technology, when other vehicles around it will not be automated.
The Galileo global navigation satellite system, in combination with other positioning and sensor technologies, is the answer to this positioning puzzle. The innovative solution was developed in the PRoPART project – Precise and Robust Positioning for Automated Road Transports – which involves Scania and six partners – and could be a key enabler for autonomous transports in the future.
‘Centimetre-level’ positioning
The solution was recently demonstrated in a recreated motorway situation at the AstaZero test area in Sweden, with a connected autonomous truck and two unconnected manned cars.
As part of the test, a Scania self-driving truck executed a safe and efficient lane change in traffic. The manoeuvre was managed by the new system, relying on centimetre-level positioning combined with collaborative perception sensor data.
The project demonstrated that it was possible to pinpoint the position with ten-centimetre accuracy. The truck could execute the manoeuvre due to the precise positioning and an accurate representation of the whole surrounding environment. This was achieved by fusing data from the truck’s camera and front and side radars combined with radars mounted on roadside units.
Infrastructure-to-vehicle communications
“In addition to positioning, we’ve also added infrastructure-to-vehicle communications,” says Project Coordinator Stefan Nord, RISE, the Swedish Research Institute.
Ordinarily, autonomous vehicles rely on their own sensors to interpret and process data on the surrounding environment. “If vehicles share information, you can extend their horizon and benefit from data from another vehicle to also look around the corner and thereby gather more data as a basis for manoeuvring decisions,” explains Nord.
The project demonstrated that it was possible to pinpoint the position with ten-centimetre accuracy. The truck could execute the manoeuvre due to the precise positioning and an accurate representation of the whole surrounding environment. This was achieved by fusing data from the truck’s camera and front and side radars combined with radars mounted on roadside units.
The new technology is a simple but profound breakthrough. The FRAS system is already a valuable resource, but Scania’s service technicians around the world use different everyday expressions that aren’t necessarily the official terminology or proper word sequence. In the case of the S 650 customer, a search in FRAS for “uneven idle” was fruitless. But an AI search was immediately successful. The indicated software update was carried out, the problem was fixed, and the customer drove away happy.
Scania Great Britain’s technical support annually receives in excess of 10,000 FRAS cases from workshops, including both technical questions and quality deviation reports.
“In most cases we respond to the workshops with answers using information that is already available in FRAS or via other media available in Scania systems,” says Technical Manager Aaron McGrath, Scania Great Britain. “By utilising the AI search at workshops we can get the information exactly where it is needed; i.e. at the technicians’ fingertips, saving on troubleshooting lead-time and also improving customer uptime.”
About PRoPART
The PRoPART project combined Real Time Kinematic positioning software from Waysure (Sweden) with satellite measurements from Fraunhofer IIS (Germany). The satellite positioning was augmented with an ultra-wideband ranging solution from Spanish research institution Ceit-IK4.
The self-driving truck was supplied by Scania, with Hungary-based V2X company Commsignia providing the short-range communication technology. Baselabs from Germany provided sensor data fusion of onboard and roadside sensors and developed a situational assessment for the intended automated lane change manoeuvre. The project was coordinated by RISE. The project has received funding from the European GNSS Agency under the European Union’s Horizon 2020 innovation programme.
