Ten cameras in different systems with a wide range of purposes, long and short-range radar systems, fusion of the resulting data and reconciliation with stored values, networking with traffic light systems and an automatic braking system – these are the technical requirements of the Mercedes-Benz Future Bus with CityPilot for semi-automated driving on BRT routes. The CityPilot is another milestone reached by Mercedes-Benz on the road to autonomous driving. The CityPilot is based on the Highway Pilot of the Mercedes-Benz Actros, however it exceeds the latter's capabilities to meet the needs of its specific area of operation: new functions include traffic light recognition, pedestrian recognition, centimetric precision when halting at bus stops and the ability to drive semi-autonomously in tunnels. In this way the bus becomes one with its environment not only with its design, but also with the technology it uses to move along its line and communicate with its surroundings.

BRT lines are ideal for autonomous driving

Always the same route on a separate line or track, a clearly defined timetable, defined and identical actions at bus stops: regular service city buses on BRT lines (BRT = Bus Rapid Transit) are ideal for autonomous driving. Both in the truck and passenger car sectors, Mercedes-Benz is the leader in taking steps on the way towards autonomous driving. Transferring this comprehensive know-how to the regular city service bus sector is therefore logical.

Bus operation is however subject to certain special circumstances – this is why the technology cannot simply be adopted from other vehicle systems, but must rather be developed further in the important aspects and where necessary also supplemented regarding the specific operating conditions. This applies to typical traffic situations such as traffic lights and pedestrian recognition, vehicles ahead in the same lane, passing through tunnels, negotiating junctions controlled by traffic lights, stopping and departing from bus stops and automatic opening and closing of passenger doors.

CityPilot – a highly specialised and unique technical feature

The specific operating conditions for a city bus therefore require equally specific technical equipment for autonomous driving – the cost and effort required for monitoring the road and the surroundings is extraordinarily high. Mercedes-Benz can however fall back on extensive experience with the Future Truck. This includes features such as long-range radar with a range of up to 200 m, electrically actuated Servotwin steering and the mirrorcams instead of exterior mirrors. Also familiar is the lane-tracking camera, which is used for the Lane Keeping Assist systems in other Mercedes-Benz buses and trucks. A further lane-tracking camera is used as an additional safeguard.

There are no less than four short-range radar sensors – two in the front section and two at the front corners – to cover distances from 50 centimetres to ten metres ahead of the bus. Two stereo cameras with a range of up to 50 metres allow 3D vision and recognition of obstacles and pedestrians.

Precise positioning with centimetric accuracy

Precise positioning of the Mercedes-Benz Future Bus is ensured by the satellite supported location system GPS, the lane-tracking cameras and four cameras for global visual location. These cameras are installed at front axle level high up on the sides, monitoring the surroundings and comparing them with images pre-stored in memory. Their purpose is to ensure exact positioning, and they are guided by waypoints. They operate to an accuracy of eight centimetres and are also used in illuminated tunnels. Such cameras were first used three years ago, for the autonomous journey of a Mercedes-Benz S-Class on the Bertha-Benz Memorial Route.

Two further close-range cameras are directed vertically downwards at the front sides. These recognise the pattern of the asphalt road surface like a fingerprint, and likewise compare this continuously with previously stored images of the route. And finally there are three cameras recording the journey. They record both the movements of the bus and the actions of the driver.

In this way a complex process of sensor fusion creates a precise picture of the local environment, with the exact position of the bus in its immediate surroundings. This means that it moves along its lane with centimetric precision. More precisely than a driver could ever hope to achieve manually in day-to-day operation.

Networking technical data with the traffic light infrastructure along the route ensures early recognition of each traffic light status, thus allowing a predictive, consistent and as a result more fuel-efficient driving style than is possible by conventional means.