Keywords

1 Introduction

Recent advances in automotive electronic control technology have led to the use of steer-by-wire systems in steering systems. Using a steer-by-wire system makes it possible to control the steering gear ratio [1]. Therefore, by setting the steering gear ratio to a high gear, it is possible to reduce the steering burden. However, the resolution of a steer-by-wire system with a high gear ratio is reduced, and there is concern that the system’s operability will deteriorate. Previously, it has been shown that if the torque is too heavy, it causes a physical burden, whereas if it is too light, it causes driving difficulties [2]. Therefore, there is also concern that inappropriate steering reaction torque may cause operating errors.

This research aims to construct a system to control steering reaction torque by evaluating parameters for each driver. However, the position of a human driving a car varies according to the physique of the upper limb in terms of joint angle, position, arm length, etc., and the moment generated at the shoulder joint differs from person to person. However, the position of a human driving a car varies according to the physique of the upper limb in terms of joint angle, position, arm length, etc., and the moment generated at the shoulder joint differs from person to person. This study proposes a load experiment in which drivers with different upper limb physiques can produce the same load at the shoulder joint. In this experiment, we confirmed that we have sufficiently replicated the burden of the steering operation.

With a high steering gear ratio, it is assumed that small driver inputs have a greater impact on the vehicle compared to a standard steering system. Therefore, the driver’s condition is a concern regarding its effects on the vehicle. To address this, we will conduct load experiments to examine how the shoulder joint’s burden varies daily.

2 Estimation of Steering Burden Using an Upper Limb Burden Model

2.1 Calculation of Shoulder Joint Moments Using the Upper Limb Burden Model

First, using Simscape, an analysis software by MathWorks, we created an upper limb burden model that reflects the driver’s upper limb physique and replicates the steering operation (Fig. 1). Previous studies have demonstrated this model to replicate steering burden sufficiently [3, 4].

The x-axis is defined as the direction perpendicular to the vehicle’s forward direction, the y-axis as the vehicle’s forward direction, and the z-axis as the vehicle’s vertical direction. The shoulder joint is denoted as the origin O, the elbow joint as Q, and the position where the steering wheel is grasped as P. A ball joint is used at the shoulder joint O and the gripping position P of the steering wheel, while a rotational joint is used at the elbow joint Q. The moments Mx, My, and Mz around the x, y, and z axes, respectively, at the shoulder joint O on the steering side, when the steering reaction force due to the steering torque applied to the steering wheel is applied at the gripping position P, with the weight of the lower arm and upper arm acting at the centers of mass R and S, respectively, were calculated using Eq. (1). The moments Mx, My, and Mz at the shoulder joint O, as shown in Fig. 1, are the moments occurring around the x, y, and z axes of the shoulder joint O in the upper limb load model. These represent the shoulder joint moments generated around the x-axis, y-axis, and z-axis, respectively.

$$ M_x = y_P \cdot F_z + y_R \cdot m_1 g + y_s \cdot m_2 g $$
(1)
Fig. 1.
figure 1

A Simscape-created upper limb burden model.

Full size image
Fig. 2.
figure 2

Appearance of steering experiment in ultra-compact EV.

Full size image

The parameters used in Eq. (1) is shown below. yP: y-coordinate of the steering wheel grip position [m], upper arm’s center of mass position [m], yR: y-coordinate of the upper arm’s center of mass position [m], yS: y-coordinate of the lower arm’s center of mass position [m], Fz: z-axis component of the steering reaction force torque [N], m1: mass of the upper arm [kg], m2: mass of the lower arm [kg], g: gravitational acceleration [m/s2],

2.2 Analysis Condition

In this study, the parameters of a 23-year-old male, as shown in Table 1, were utilized to reproduce the driving position of the experimental participant when riding in the ultra-compact EV (Fig. 2). The steering reaction torque was set to 0, 1.0, 2.0, 3.0, 4.0 [N･m]. The steering operation is to turn the steering wheel from 0° to 90° in the left direction. The shoulder joint can evaluate the steering burden in the opposite side of the steering direction [5, 6]. Therefore, the maximum shoulder joint moment on the right side is calculated for each steering reaction torque. Figure 3 shows the shoulder joint moment Mx calculated at each steering reaction torque. The results show that shoulder joint moments occur in the range of 10 ~ 23 [N･m].

Table 1. Parameter of a participant.
Full size table
Fig. 3.
figure 3

The Result of the shoulder joint moment in steering operation.

Full size image
Fig. 4.
figure 4

The method of load experiment.

Full size image

3 Variation in the Amount of Steering Burden Caused by the Driver’s Condition

3.1 Reproduction of Steering Burden by Load Experiment

The load experiment proposed in this paper can be applied mostly to the anterior deltoid muscle. In addition, the muscle activity of the anterior deltoid is known as a muscle that can be used to evaluate steering burden [5, 6]. Surface EMG measures muscle activity. The location of the electrodes for measuring the surface EMG of the anterior deltoid muscle is shown in Fig. 4 (a). The surface EMG was amplified using a Bio Amp ML132 (AD Instruments), passed through an A/D converter PowerLab ML825 2125 (AD Instruments), and recorded at a sampling frequency of 4 kHz. The original measured waveform was smoothed using the root mean square (RMS) every 0.1 s. A comparison of muscle activity between dynamic steering operations and load experiments was conducted as a preliminary experiment. The results confirm that the burden is matched within 10% of the %MVC of the anterior deltoid muscle.

Table 2. Calculated weight from shoulder joint moment.
Full size table
Fig. 5.
figure 5

.

Full size image

The shoulder joint moment Mx calculated by the upper limb burden model during the steering operation is reproduced in the load experiment. Assuming a load experiment as shown in Fig. 4 (b), the weight is calculated such that the shoulder joint moment during steering can be calculated using Eq. (1). The calculated results are shown in Table 2. The range of shoulder joint moments was also set smaller than the range during steering operation. In addition, a negative weight is when the weight exceeds the arm’s weight. Force is applied upward (opposite direction in Fig. 4 (b)).

A load experiment uses weights corresponding to the respective steering reaction torques. Teach the participant to hold the weights and measure the surface EMG potentials for 5 s. This will be done five times for each weight. This experiment was conducted for five consecutive days to evaluate the driver’s daily burden. The surface EMG is evaluated using %MVC, expressed as a percentage of the MVC (Maximum Voluntary Contraction) at maximum effort.

3.2 Experiment Result and Consideration

The average daily %MVC measured for 5 s is shown in Fig. 5. As the shoulder joint moment increased, the load on the shoulder joint increased, resulting in a trend toward an increase in %MVC. When the shoulder joint moment was at its lowest, the difference was about 3% on different days. However, the greater the load on the shoulder joint, the greater the difference in %MVC. At 23 [N·m], the difference increased to over 10%. An increase of 10% in load is equivalent to an increase from 14 [N·m] to 23 [N·m] on Day 1. This difference in shoulder joint moment is equivalent to the load difference when the steering reaction torque in Fig. 3 increases from 1.0 [N·m] to 4.0 [N·m]. Therefore, depending on the driver’s condition, the burden due to the steering reaction torque may vary significantly.

4 Conclusion

If drivers are evaluated by their steering operation, the amount of steering burden cannot be normalized for each driver. Therefore, we proposed a load experiment that can normalize the steering burden for each driver. The load experiment subjected drivers to the same steering burden for five consecutive days, and the difference in load increased as the burden increased. A load experiment subjected the driver to the same steering burden for five consecutive days.

As a result, the difference in burden increased as the load increased. The load experiments also revealed that the steering reaction torque is felt very differently by drivers under different conditions. We also showed that the methods in this paper can be used to evaluate the effect of driver condition on the amount of burden.

However, this paper has not been able to address the causes of driver conditions. In the future, we are also going to evaluate the driver’s fatigue level at the same time to examine what conditions affect the steering burden.