Highly automated driving is one of the most important innovations in the automotive industry. Many OEMs, component suppliers, IT companies and startups are working on solutions that are translating automated driving into products. IAV is helping its customers with innovative concepts and methods to make safe and comfortable highly automated driving some- thing that the consumer can soon experience too.
Whatever happens, just don’t cause an accident! This maxim applies much more to highly automated or autonomous driving than it does to human drivers. But what happens if excessive caution makes the safety functions hit the brakes at the slightest cause for concern and constantly lunge occupants forward behind their seat belts? “Highly automated driving also needs to provide ride comfort”, says Benedikt Schonlau, head of the Active Safety department at IAV. “Otherwise consumers won’t accept it. For this reason: The higher the speed, the more important it is to distinguish between really critical and harmless traffic situations.”
Introducing highly automated vehicles into volume production places two central demands on vehicle safety. To begin with, it must be remembered that technical faults are forgiven far less than mistakes by a human being behind the wheel. And highly automated vehicles should not cause accidents in situations in which human drivers typically do not cause accidents as this would reduce acceptance of this technology.
Slowing down comfortably or emergency stop?
To meet these demands, IAV is pursuing a multiple-layer function architecture for the automated driving functions. The main aspect of this function architecture is to satisfy the demands on ride comfort in one layer, with the zero-accident guarantee being implemented in another layer (active safety layer). “If a critical driving situation should occur, the active safety layer takes over command and returns the vehicle to a safe state”, Schonlau explains. The active safety layer works with far fewer demands on ride comfort, also reducing overall complexity. IAV’s developers presented this approach for the first time at the 2015 FAST-zero symposium in Gothenburg with a demonstration and various presentations.
This architecture also makes it possible to implement various methods for developing the different layers. For instance, an agile methodology is used for the comfort layer as a way of being able to respond quickly to changing consumer demands and experience from the field. A waterfall approach is used for the active-safety layer to meet the safety objectives in the product top down.
Driving-environment sensors, V2X and cloud-based signals
Alongside the system architecture, the availability of data is a key challenge when it comes to getting highly automated vehicles into volume production. Three principal core technologies can be applied. For the driving-environment sensor system, it is important to devise a clever combination of sensor technologies such as radar, laser, camera and ultrasound. Vehicle-2-X communication provides the capability of reliably delivering dynamic information in real time. One example of this is the current signal color of traffic lights. Ultimately, cloud-based signals also support data recorded by the sensors of many different road users, aggregated in a backend from many signals and then communicated back to individual vehicles on the basis of their position.
“Cloud-based signals in particular provide a previously underestimated source in relation to added safety and ride comfort for highly automated vehicles. However, it is not possible to say when these data can be used for any heavy intervention into driving dynamics as this presents problems from the aspect of security and also in terms of product safety”, Schonlau explains. However, even today, the vehicle can be preconditioned for specific situations, such as for a crowd of pedestrians near a driving path by reducing speed. This makes it possible in an acute danger situation to reduce the hazard with far less intervention. This approach enhances ride comfort and indirectly improves road safety by a better flow of traffic and by responding sooner to hazards”, Schonlau says.
In future, this particular application will be supported by wearable computers. They can be fitted to garments, cyclist helmets or to bicycles and record and process data via sensors. IAV’s developers are working on getting the cloud to use wearables for identifying the swarm behaviour of pedestrians and taking this as the basis for issuing driving recommendations. The vehicle will then decide how to turn the recommendation into practice. This use case was shown in a joint presentation by IAV and Microsoft at CES 2016 in Las Vegas. “If these technical developments enhance road safety, it will also improve acceptance among the population”, Schonlau says.