TASCC: Human Interaction: Designing Autonomy in Vehicles (HI:DAV)

Cars that can drive themselves have been predicted for some time, but they are nearly with us. Highly automated vehicles are likely to be on public roads within the next ten years. The largest gap in our understanding of vehicle automation is how drivers will react to this new technology and how best to design the driver-automation interaction. Our project will answer these questions by studying a wide range of drivers with different driving experience in simulators, or test-tracks and in road going vehicles. We will start by modeling driver behaviour in our laboratories in order to help us design inclusive, user-centered, interfaces with vehicle automation. We plan to learn from different situations where automation is already used (such as the Boeing versus Airbus design approach of 'soft' and 'hard' automation). Then we will test the designs out in our driving simulator (which comprises a Jaguar XJ connected to computers with large projectors and screens). We will test drivers of different ages, gender, experience and capabilities, in a range of scenarios (e.g., different road types and environmental conditions) with different automation systems (e.g., autonomous driving, auto 'valet' parking, adaptive vehicle personalization, off-road assistance) and different interfaces (which we will design). Our design approach will personalise the driver interfaces to the widest range of drivers possible (e.g., wide ranges of age and driving experience). We want the driver-automation interfaces to be intuitive and the necessary responses to be obvious. We design so that the system adapts to you, not you having to adapt to the system. We will also monitor driver behaviour over successive weeks of use. This will help us to understand how drivers learn to respond to the automation over time. We will be studying how control of the vehicle is handed back and forth between the automation and the driver. We are also interested to know what the driver is doing when the vehicle is under automated control. To help in this understanding, we will be monitoring the driver physiological and psychological states. The idea is to use this information to adapt the automation to the driver and the situation, so that performance of the system is optimised. This will enhance safety to the benefit of the driver and other road users. The studies will progress from the simulator to the test-track, as our interaction and interface designs evolve with testing. On the test-track we can record driver behaviour physiological and psychological states to see what further changes are needed and whether the automation can be even more highly tailored to the situation. As the research progresses we will take the revised designs into road going vehicles for the final set of tests. These tests will be used to validate the designs and prepare them for delivery into production vehicles. At the end of the research, we will be able to provide JLR with the design methods and well as the designs themselves. This means that they will be able use the methods we have provided to design new systems into the future. During the course of the research the Universities of Cambridge and Southampton will be working closely with JLR engineers to ensure that the UK remains at the forefront of technological innovation in vehicle automation. We will have answered the questions about how drivers will react to this new technology and how best to design the driver-automation interaction. The success of vehicle automation design will be on designing appropriate interactions and interfaces that support the driver. Our research will be essential to that success.