ARCOS investigated four driving functions –time headway control, early warning on downstream incidents, collision avoidance, and lane departure avoidance (Blosseville et al., 2003). Car manufacturers have already introduced some devices in the car in order to contribute to safety with regard to these functions. For example, ACC (Adaptive Cruise Control) consists in an automatic control of time headway; ABS (Antilock Braking System) is related to collision avoidance preventing from skids; ESP (Electronic Stability Program) deals with lane departure avoidance preventing from spins. One of the main ARCOS research objectives was the introduction of a human-machine cooperation point of view leading to enlarge the consideration of a particular device. As a matter of fact, a device can be used within various human-machine cooperation modes, which are a re-interpretation of levels of automation in terms of cooperation. Four modes were defined in relation to the way the device cooperates with the driver. The perception mode is an enhancement of the driver perception (e.g., HUD enhancing vision in fog). It relies on the infra-red perception capability of the device. The mutual control mode provides the drivers with a criticism of their behaviours (e.g., a warning when there is an impending lane departure). The device know-how is concerned. The function delegation mode provides the drivers with a partial automation of their driving tasks (e.g., time headway control, letting the lateral control –steering– to the driver). The full automatic mode can be used in very hazardous situations (e.g., emergency braking for collision avoidance or mitigation).

The three first contributions of this special issue are strongly related to human-machine cooperation. Two of them are dealing with the delegation of time headway control. Rajaonah, Anceaux, and Vienne address a very general problem of human-machine cooperation –the driver trust in the human-machine system. This trust is decomposed into three components –trust in the device, self-confidence, and trust in the cooperation with the device. However, the last one seems to result from a combination of the two first. Tricot, Rajaonah, Popieul, and Millot explore the possible benefit of more adaptive ACC than the current ones, including adaptation to both driving style (e.g., preference in terms of braking strength) and environmental conditions (e.g., incoming bend in the road). The new device has a positive effect on the device rate of use, which is also influenced by ACC use experience. The third contribution focuses on the complementary function –lateral control. Hoc et al. evaluate a wider variety of human-cooperation modes. Delegation of lateral control produces difficulty in returning to manual control when the device is invalid (obstacle skirting). A warning (auditory and haptic) mode produces a positive effect, whereas the effect of an action suggestion mode (haptic on the steering wheel) is not conclusive (high individual differences).

The last contribution in this issue, as well as the contribution that will be published in the next one, are devoted to the mechanism of human perception in collision anticipation and avoidance. Meskali, Barbet, Espié, and Bootsma show that braking appears strongly related to the perception of the optic expansion of the obstacle. Barbet, Meskali, Berthelon, Mottet, and Bootsma find that driving experience does not affect this mechanism, but aging, which results in a modification of the perceptive mechanism when the environmental structure is poor.

This issue stresses two main research and application orientations for the future. First, a correct calibration of human-machine cooperation should be an important concern, in order to take the best advantage of the two driving entities. Second, more attention should be devoted to sensorimotor models for better interventions of the technical devices into the human control loops.