Partially- and highly automated driving in a military platoons

Partially and highly automated vehicles integrate assistance and automation systems that they can be deployed at different assistance and automation levels. In suitable situations, the driver is therefore able to temporarily delegate the driving task to the automation system. A crucial quality of these systems however, is the cooperation between the driver and the automation and that the driver can resume control at any time and in any automation mode.

The »Highly automated platooning of ground vehicles (StrAsRob)« project illustrates this concept with an automated following function for military trucks as an example application and aims to demonstrate the capabilities of this system with professional military drivers in a simulator.

The concept of cooperative vehicle guidance is based on a generic approach to describe the degrees of freedom of interaction between humans and automation. Two existing implementations of cooperative vehicle guidance are conduct-by-wire (CbW), which operates by discretely assigning maneuvers, and the H-mode. The latter enables a continuous and direct interaction via the traditional control elements such as the steering wheel and gas pedal or via alternative control elements such as active sidesticks and contains discrete components on the maneuver and control level. Numerous complex technical functions are integrated in such a way that they can be provided to humans as a coherent, intuitively operated co-system for partially- and highly automated vehicle guidance.

The basis of this system is a four-level assistance and automation scale:

AAM1 (»driver only« to »assisted«)
The vehicle is controlled largely independently by the driver. However, the automation system gives suggestions for action, for example in the form of haptic signals via the control unit.

AAM2 (»assisted«)
Active assumption of the longitudinal guidance by the Adaptive Cruise Control (ACC) controlled by the automation system and passive assistance that informs the driver.

AAM3 (»partially automated« to »highly automated«)
The operator can trigger situation-dependent driving maneuvers with haptic gestures on the control unit. These maneuvers are carried out independently by the automation system and can be felt, modified, changed and aborted by the operator at any time. The driver thus not only retains control, but also remains actively in the loop.

AAM4 (»highly automated« to »fully automated«)
Under appropriate circumstances, highly automated driving is implemented, where the driver can also leave the vehicle's control circuit for a limited time.

The »StrAsRob« project transferred the fundamental ideas of cooperative vehicle guidance and the H-mode to the interaction with highly automated trucks. The project kicked off with a number of workshops with various stakeholder groups (drivers, vehicle registration office, designers and developers) to define potential use cases. Typical examples of those are setting up the vehicle, assembling the platoon, highly automated or assisted following of the leader vehicle or temporary disbanding of the platoon. The four modes of the assistance and automation scale were applied to these tasks, and an interaction concept was developed. To test usability, three pre-set sequences were designed and evaluated in the simulator.

The Base Kata scenario was a drive in a cooperative and highly automated platoon. The trial participants were 13 Bundeswehr drivers. In the simulation, they were seated in the second vehicle of the platoon and were followed by another truck. They evaluated the overall system through a questionnaire at regular intervals for the individual use cases.

The individual test kata were evaluated by the trial participants following three drives:

1. Naive drive
2. Briefed drive
3. Trained drive

Results: In spite being at an early stage in the development, the initial evaluation was very positive. It became clear that the drive following the thorough training resulted in an even better overall evaluation of the entire system. When comparing the assessments across the »Naïve«, »Briefed« and »Trained« drives, the differences in the transition from the automation level "loose rein" (partially / highly automated, driver »in the loop«) to »secured rein« (highly / temporarily fully automated, driver »out of the loop«) are particularly striking. This suggests that the differences between the levels were not self-explanatory enough, a clear indication where potential improvements can be made to the system. The integrated function of reassumption of control by steering wheel actuation also holds potential for improvement, as it was not taken full advantage of, even after explanation and training.

As networking and autonomous capabilities increase, vehicle cooperation is becoming increasingly more important – not only between the automation system and the driver, but also between the different vehicles and between vehicles and infrastructure. Against this backdrop, the »StrAsRob« project is a concrete and promising example of cooperative vehicle guidance that should be moved towards series development.

• Fraunhofer Institute for Communication, Information Processing and Ergonomics FKIE
• Federal Office of Bundeswehr Equipment, Information Technology and In-Service Support (BAAINBw)
• Diehl Defence GmbH & Co. KG
• Rheinmetall AG
• Bundeswehr University Munich