Keywords

1 Introduction

The knowledge of the forces applied by the driver to the steering wheel is of great utility for assessing the driver-vehicle interaction, either for normal or emergency driving manoeuvres. Such a knowledge is crucial for developing new driver models based on Neuro-Muscular System (NMS) activation [1]. NMS can be activated consciously or unconsciously, if a kick or a shock is applied to the vehicle. Improving ADAS requires the knowledge of forces applied by the hands at the steering wheel [2]. During normal driving, the driver controls the steering wheel by applying forces and moments in all the directions and not only tangentially, the investigation of such a fact has been attempted in the literature. Simple torque sensors have been successfully employed for a long time for measuring the torque applied to the steering wheel and developing the control logic of Electric Power Steering (EPS) systems. In [3], an ISW able to measure the three forces and moments was presented. Such a device was made from a normal steering wheel divided into three sectors. The ISW was tested on a driving simulator. In this paper, an innovative ISW has been used to measure the forces and the moments applied by the driver during steering manoeuvres. The steering wheel has been designed for the application on real vehicles. The mass, the stiffness, the eigenfrequencies, the power system have been defined for the use on a passenger vehicle. In Sect. 2, the ISW is shown and a brief description of its components is provided, in particular the six axis load cells and the grip force detection system. In Sect. 3, the forces and moments applied by each hand on the ISW have been considered to analyse the behaviour of different drivers in certain driving scenarios. The common attitudes and behaviours of the drivers have been highlighted and analysed. Two cases have been considered, namely conscious and unconscious steering actions.

2 Experimental Setup

2.1 Instrumented Steering Wheel

An ISW which measures the six forces and six moments applied by the driver is presented [2]. The ISW, shown in Fig. 1, has a carbon fiber composite body, two six-axis load cells to measure forces and moments, two handles that the driver must grasp and an electronic-box for signal conditioning. Furthermore, by six mono-axial load cells positioned in the handles, it is possible to measure the grip forces at each hand.

The force and moment signals are sent to an on-board vehicle acquisition system which stores the data, together with the vehicle dynamics data published on the CAN network. Due to the fast transient manoeuvres in which the steering wheel has been used, the compensation of the static weight and of the inertial contribution due to the vehicle dynamics [2] and to the steering rotation is necessary.

Fig. 1.
figure 1

Instrumented Steering Wheel

2.2 Test Setup

The kinematics of the vehicle were acquired with an OxTS inertial measurement unit, capable of measuring vehicle accelerations in three directions, together with roll, pitch, yaw angles and their derivatives. A GPS antenna was used to track the path of the vehicle. Speed and steering angle data were obtained from the vehicle’s CAN network.

Conscious Steering Action: Sine Test.

In the sine test the task of the driver is to steer the vehicle at a fixed frequency and fixed amplitude of steering wheel angle oscillation. This sine input is chosen as a high-consistency method, due to its “open-loop” characteristic.

Conscious Steering Action: Handling Circuit Test.

The drivers are asked to follow a reference trajectory as close as possible, while driving at high speed. The reference trajectory is a white solid line at the inner edge of the racetrack’s width, and it does not correspond to a racing line.

Unconscious Steering Action.

The vehicle is driven in a kick-plate test, where a moving platform introduces a lateral sliding to the rear axle of the vehicle when it passes over it. In this way, an emergency situation is simulated in which the driver suffers an unexpected lateral disturbance and must keep the vehicle in trajectory by acting on the steering.

3 Data Analysis

3.1 Conscious Steering Action

Sine Test Results.

When asked to perform the sine test manoeuvre, different drivers show different approaches when comparing tangential forces applied on the steering wheel. Considering that tangential forces are the main contributors to the steering action, Fig. 2 shows that the more experienced driver (in green) uses a balanced approach, with the two hands mostly cooperating during the steering action. The other two drivers show a “4th quadrant” behaviour, which means that the tangential forces do not cooperate but rather fight each other.

Fig. 2.
figure 2

Tangential Forces in sine test, comparison of 3 different drivers

In other words, less experienced drivers prefer to pull downwards with the inner hand rather than pushing upwards with the outer hand and as a result, the force applied by the inner hand has a higher absolute value for them with respect to the professional driver.

Handling Circuit Test Results.

In this test the same pattern of dominance of the inner hand is seen once again, as shown in Fig. 3. In this example, as drivers turn the steering wheel clockwise, the tangential force exerted by the inner hand (right one) increases significantly, while the left hand exerts a smaller force in the opposite direction (i.e. without cooperation). In this test all drivers showed a similar behaviour, regardless of skill level.

Fig. 3.
figure 3

Tangential forces during a left-to-right chicane, dominant inner hand of different drivers

Steering Quality Index.

The classification of the drivers’ behaviour is based on heuristic algorithms [4, 5]. For each driving scenario, various analyses have been completed and, when possible, some Key Performance Indicators (KPI) have been calculated to reflect the driving perception. A KPI for the correlation between objective performance of the vehicle and subjective evaluation of the drivers has been defined as follows [4]: The steering quality index is the standard deviation of the intersection points between the sum of the resultant forces for both hands and a plane perpendicular to the average sum of resultant forces, set at a fixed distance of 0.5 [m]. This index has been measured during the end of the steer-in phase, where the steering wheel torque |SWT| has a local minimum. The idea behind this objective indicator is that a subjectively bad steer-in manoeuvre is characterized by a high jerkiness of forces at the end of this phase. This jerkiness is reflected both by a change in direction of the forces and by a change in amplitude, the former being the main contributor.

In Table 1, the steering quality index has been computed for 5 vehicle setups used by the professional driver. Vehicle 2 has been used as a benchmark for its well-known good steering characteristics. The steering quality index correctly ranks the benchmark vehicle as the better one.

Table 1. Steering quality index

3.2 Unconscious Steering Action

The ISW is used to assess the unconscious application of forces and moments at the steering wheel during a kick-plate pass-by manoeuvre. The vehicle passes over a plate that is kicked sideways to cause a spin. A prompt counter-steering action is requested to the driver. Due to the side kick, an unconscious steering action is generated, due to the activation of the Neuro-Muscular System.

Fig. 4.
figure 4

Unconscious activation of NMS and steering action due to side kick applied.

In Fig. 4, the unconscious activation of the NMS is shown. Actually, at t = 53.7 s the vehicle is kicked, as is possible to note from the lateral acceleration graph. Then, at t = 53.87 s the unconscious reflex action reaches its maximum to counter the undesired vehicle motion. At t = 54.2 s the conscious steering action is initiated. The unconscious torque is 20% of the absolute value of the maximum torque during the recovering manoeuvre. Analysing a panel of nine different drivers, the maximum of the reflex torque occurs at 0.13 ± 0.03 s after the vehicle disturbance and at 0.28 ± 0.11 s before of the actual steering. Moreover, the unconscious moment applied by the driver influences the vehicle dynamics, e.g. the yaw acceleration. Such an effect highlights the need of a proper modelling of the unconscious driver action, because of its influence on the vehicle dynamics.

4 Conclusion

The presented ISW appears to be the most accurate device for detecting forces and moments applied to the steering wheel by each hand of the driver. The ISW has been used to characterize driving patterns of both professional and regular drivers. Relevant differences exist between forces and moments applied by regular and professional drivers during the sine wave test or the handling circuit test: the professional driver is more efficient in using tangential forces to perform the steering action. The steady state turning analysis has also been considered for analyzing the distribution of tangential forces that is different at the two arms. These forces distribution looks quite consistent between different drivers. A steering quality index has been computed to correlate the measurements of the ISW with the subjective evaluation of the professional driver. Further validation of the index could focus on more repeatable manoeuvres such as U-turns, lane changes and 90° corners as well as on a robust detection of the steer-in phase. Referring to unconscious manoeuvres, occurring when a kick or shock is applied to the vehicle, the forces and moments due to the muscular unconscious activation are considerable, up to 20% of the maximum ones. The muscular activation occurs tenths of second before of the actual steering of the steering wheel, influencing the vehicle dynamics and highlighting the need to model the unconscious reflex muscular activation in the NMS driver models. Such an information could be helpful in designing innovative ADAS (Advanced Driver Assistance Systems) that can intervene in much less time that the actual control systems, increasing the road safety.