Continental Achieves Highly-Automated Driving Using Production-Ready Technologies

Continental, the international automotive supplier, will today be presenting an Architecture Migration Demonstrator at the EU-sponsored HAVEit research project's final event; the demonstrator is equipped with production-ready technologies which allow highly-automated driving. Using a co-pilot system, the automated demonstrator can maintain a set speed or the distance from the vehicle in front and keep to its lane while, at the same time, using a camera to carefully monitor how attentive the driver is to the traffic situation. The software developed as part of the HAVEit project could soon, i.e. within about five years, be developed into a production-ready, fully-integrated system covering the initial automated driving functions, said Holger Zeng, Continental's HAVEit project manager.

Focus on architecture and implementation

Producing a common HAVEit system architecture and bringing it close to production readiness were the main tasks when developing the Architecture Migration Demonstrator. The HAVEit system architecture was first worked out in detail top-down and clearly specified, followed by implementation, using the AUTOSAR methodology and the CESSAR configuration system. Continental's automotive control unit, the Chassis & Safety Controller (CSC), acted as the universal basic computer on which all the HAVEit system's sub-systems were programmed. Several of these CSCs are networked with standard CAN buses so that highly-automated driving can be achieved with inexpensive, currently available technology. Continental's CSC is also installed as an integral element of the HAVEit system in all the other HAVEit demonstration vehicles; so, they too are making use of a piece of series production technology. Other Continental components used are the long-range ARS 300 radar sensor for object detection, the electronically controllable braking system and electric power steering (EPS). The vehicle is also fitted with two Continental cameras: one keeps an eye on the road up to 60 meters ahead of the vehicle and recognizes traffic lanes; the other monitors the driver.

The data from the sensors is analyzed using data fusion, which enables the co-pilot system to achieve the automation functions. At today's HAVEit final event, these basic automation functions will be demonstrated using a motorway-type traffic situation. In the partially-automated mode, the system operates in the longitudinal direction like an adaptive cruise control system; the highly-automated mode provides an additional lane-keeping capability. During highly-automated driving, the onboard camera keeps a permanent watch on the driver, including observing where he is looking and how often he blinks; if the driver's attention wanders, he will be obliged to resume control of his vehicle. The system will also hand back control to the driver if conditions no longer make highly-automated driving possible, e.g. if no road markings can be detected or if the bends are too tight. If the driver fails to react, the vehicle's speed will be gradually reduced until the vehicle comes to a standstill.

Although more complex extreme scenarios, such as lane changing, driving around obstacles and emergency braking, have not yet been implemented in this demonstration vehicle, it will nonetheless be technically possible to achieve them in future using the software which has already been developed, together with an upgraded system for monitoring the vehicle's immediate surroundings; this has already been demonstrated by some of the other HAVEit vehicles. This car proves that we have made the gap between research and product development smaller than ever before, Zeng added. As for the future, using the latest generation of the Chassis Domain Control Unit (CDCU) rather than the CSC as the universal basic computer offers the potential for either improving the vehicle's performance still further or alternatively for reducing cost and engineering effort. The research vehicle was developed at Continental's Regensburg site, assisted by our HAVEit project partners.