Development of a driving assist system for ultra-compact electric vehicles using a steer-by-wire system: Fundamental consideration of muscle burden and evaluation by a continuous steering operation

Abstract

Ultra-compact electric vehicles has excellent environmental performance and are extremely convenient for short-distance travel. However, owing to cabin space limitations, it is difficult to mount power steering. Therefore, there is a need to increase the gear ratio of the rack and pinion to change steering angle because such vehicles need light torque to steer. However, increasing the gear ratio requires more rotations of the steering wheel. Our research group focused on developing a steer-by-wire system (SBWS) that freely controls the steering torque. Although we evaluated the burden when a driver rotates the steering wheel in one direction in a previous study. This study assumed the actual steering operation in an SBWS. And then we evaluate muscle burden when a driver steers with continuous changing of the steering direction.

Keywords

Steer-by-wire system ultra-compact electric vehicle self-aligning torque surface electromyography (EMG)muscle burden on driver

1. Introduction

In Japan, 70% of vehicles travel less than 10 km per trip, and 80% of trips convey less than 2 passengers; therefore, attention is focused on ultra-compact electric vehicles (EVs) that are extremely convenient for short-distance travel. Because the vehicle has a short wheelbase that accommodates one or two people, it performs well in small-radius turns. In addition, ultra-compact EVs have already been released by several automakers, and are used for delivery and car-sharing services in tourist areas. Ultra-compact EVs are considered an appropriate vehicle for elderly people who live in mountainous areas where transportation is essential. In short, ultra-compact EVs are likely to be used by a wide variety of people for a multitude of purposes. However, there is drawback to the steering system because of its small size. Most vehicles are not equipped with an assistive system for steering, such as power steering. Ultra-compact EVs without power steering need to increase the gear ratio of the rack and pinion that converts the rotary motion of the steering wheel into rectilinear motion to change steering angle because it needs only light torque. But increasing the gear ratio requires more rotations of the steering wheel. There is a concern that the burden of steering may lead to dangerous driving behaviors, especially among the elderly [1]. Therefore, we have proposed to equip the ultra-compact EVs with steer-by-wire systems (SBWS) that have no mechanical connection between the rack and pinion and the steering wheel (Fig. 1). In SBWS, the front wheels and the steering wheel communicate via electrical signals. Therefore, the position of the steering wheel can be set freely. In addition, an appropriate steering reaction torque can be tailored to the driver. Our research group measured surface electromyography (EMG) of the upper limb when holding a steering wheel during turning actions using an experimental device that assumes SBWS, and evaluated the muscle burden at various steering reaction torque values. We have also confirmed that tendencies of surface EMG are similar to subjective evaluations using questionnaires [2]. We demonstrate that muscle burden can be reduced by changing the steering torque pattern according to the steering wheel angle when the driver rotates the steering wheel from 0 deg. to 90 deg. in spite of subjects with different physiques [3,4]. Although we evaluated the burden when a driver rotates the steering wheel in one direction in a previous study, this study assumed the actual steering operation in an SBWS. Therefore, we evaluate muscle burden when a driver steers with continuous changing of the steering direction. The driver steers the experimental vehicle (Fig. 2) on a course where the steering direction changes continuously, simulating actual driving, and the muscle burden during steering is evaluated from the surface EMG of the driver.

Fig. 1.

Schematic diagram of the steer-by-wire system.

Fig. 2.

Photograph of the ultra-compact EV “COMS.”

Table 1

Specifications of COMS (TOYOTA Auto Body)

Minimum turning radius	3.2 m
Steering gear ratio	1:23.8
Riding capacity	One person
Maximum velocity	60 km/h
Maximum output	5 kW

2. Reaction torque generated in ultra-compact EVs

A vehicle with a normal steering system is mechanically linked. When the vehicle turns, a cornering force is generated on the front wheels, and torque is generated around the center of the installed tire. This torque is called self-aligning torque (SAT) and is applied in a direction that reduces tire side slip [5]. This is transmitted to the steering wheel via the steering shaft and becomes the reaction torque against the steering.

Fig. 3.

Position of the tie rod ends.

Fig. 4.

Photograph of the strain gauge.

Fig. 5.

Schematic diagram of the experimental apparatus.

2.1. Method of measuring self-aligning torque with a strain gauge

To know the SAT characteristics of the experimental vehicle used, we measured the SAT generated in a COMS vehicle, which is an ultra-compact EV manufactured by Toyota Auto Body. Table 1 shows the specifications of the COMS model. To measure the SAT, it is necessary to measure the axial force generated at the left and right tie rod ends (cylindrical parts that transmit steering operation to the wheels), as shown in Fig. 3 [6]. The axial force was measured from the strain on the tie rod end by attaching strain gauges (Fig. 4) on the left and right tie rod ends of the ultra-compact EV. The strain on the tie rod end is amplified by a strain measurement unit (ST-04 KEYENCE) from a single-axis three-wire strain gauge (KFGS-5-120-C1-11 Kyowa Denki) and the data logger (NR-600 KEYENCE) at a sampling frequency of 1 kHz. Since strain gauges are easily affected by noise, noise was compensated by the moving average. Figure 5(a) shows the calculation of SAT from strain on the tie rod ends. First, according to Hooke’s law, the strain is converted to stress by Young’s modulus of the tie rod end (iron). Next, the axial force is calculated by multiplying. At this time, the axial forces calculated on the left and right sides are processed, as shown in Fig. 5(b) and summed. The SAT is calculated by multiplying the total axial force by the distance from the tie rod to the kingpin. The experimental vehicle traveled in a steady circle (Fig. 6), turning at the minimum circle radius at three speeds: 5, 7, and 10 km/h. The number of laps was 5, and the SAT was measured in 15 s except acceleration and deceleration.

Fig. 6.

Traveling line.

Fig. 7.

Results of self-aligning torque measurements.

Fig. 8.

Traveling course.

Fig. 9.

SBWS device.

2.2. Results of self-aligning torque measurements

Figure 7 shows the results of maximum SAT obtained from the strain gauges. Increasing the vehicle speed from 7 km/h to 10 km/h produces an increase of more than 16.1 N ⋅ m. Considering the steering gear ratio of the experimental vehicle, the torque generated at the steering wheel is 2.268 N ⋅ m and 1.589 N ⋅ m at 7, 10 km/h. This is because as the vehicle velocity increases, the centripetal force necessary for turning increases, and the SAT increases. SAT was not measured at 5 km/h. This is likely because the experimental vehicle was turning almost statically at that speed, and the slip angle that causes SAT was approximately zero.

3. Driving experiment assumed actual driving situation

Next, the surface EMG was measured to quantitate the burden on the upper limb muscles of the driver. To compare the relative burden between a conventional steering system and an SBWS, the driving course (Fig. 8) was set assuming the upper limb muscle burden when the driver used an ultra-compact EV.

3.1. Driving course and experimental conditions

In this study, the driver was instructed to drive on the course at a steady pace of 5 km/h, based on the results of the previous chapter, to evaluate the burden from the steering operation only. In the experimental device using SBWS (Fig. 9), the reaction torque was set to generate 2.8 N ⋅ m during the operation of the steering wheel at 5 km/h, and then the driver performed the steering simulation on the driving course.

3.2. Evaluation of the burden on the upper limb muscles of the driver by surface EMG

To measure the burden on the driver, we measured the surface EMG of the anterior deltoid muscle (Fig. 10), which is the muscle that is actively engaged during steering [2]. When surface EMG is used as a means for operation analysis, the average amplitude of EMG signals is the target measurement [7]. The root-mean-square (RMS) of a section (0.1 s) was calculated at intervals of 0.1 s for the measured EMG waveform. Because there are individual differences in muscle exertion, it is necessary to normalize the acquired data to evaluate the degree of burden using EMG. We measured the surface EMG of the anterior deltoid muscle when the muscle was engaged with maximum effort [8]. The surface EMG at this time is called the maximum voluntary contraction (MVC). In the steering operation experiment, the obtained RMS was normalized by its ratio (% MVC) to MVC. The higher the value % MVC, the more burden is being placed on the muscle [9].

Fig. 10.

Photograph of position of electrodes.

Fig. 11.

Results of % MVC.

3.3. Experiment results

Figure 11 shows the results of the average % MVC for three subjects. In these data, the right direction generates positive values, and the left direction generates negative values. When the circuit is driven in the clockwise direction, as shown in Fig. 11(a), the driver steers the steering wheel to the right, so the left arm is raised, and the % MVC of the left side anterior deltoid is increased, the right arm is lowered, and the % MVC remains approximately 5%. However, when the steering wheel is turned back at 6 to 9 seconds, the % MVC of the right arm increases, and the % MVC of the left arm decreases; the magnitudes are reversed. When the steering wheel returns from −135 deg. (left direction) to 0 deg. at 9 to 12 seconds, the % MVC of the left arm increases to approximately 10%, and the % MVC of the right arm decreases to approximately 5%, but when steering wheel is maintained at 0 deg., neither value changes. When steering to the right at 22 s, the % MVC of the left arm increases, and then from 24 s to the circuit completion the driver begins holding the steering wheel angle at 135 deg., and the % MVC decreases to approximately 10%. On this circuit, the same patterns of % MVC are obtained with both the conventional steering system and the SBWS. In addition, similar results are obtained when the course is driven in the opposite direction, as shown in Fig. 11(b).

4. Conclusion

To evaluate the muscle burden in an actual steering operation, we conducted an experiment in which the driver steers continuously changing the steering direction. First, to know the characteristics of the SAT generated in the experimental vehicle, the SAT was measured by traveling in a steady circle, and the relationship between the velocity and the SAT in the experimental vehicle was obtained. Next, to evaluate the burden on the muscle of the driver that occurs when steering the SBWS and the conventional steering system, drivers navigated a course in which it was necessary to rotate the steering wheel in both directions in an SBWS device. The results confirmed that the burden on the muscle of the driver in an SBWS is same as in a vehicle with a conventional steering system. These results suggest that it is possible to examine the control of the steering torque without a traveling vehicle. It became clear that any variation in the burden on the muscle of the driver can be evaluated statically in an experimental device with an SBWS. In the future, taking advantage of the SBWS, we will measure variation in muscle burden depending on the driver’s posture, and evaluate the steering operation burden in response to various patterns and levels of reaction torque. We are also planning to investigate SBWS that can reduce the operational burden and improve operability.

References

Tamida

Nihei

Shino

and Kamata

, An analysis of the elderly on their driving behavior based on the multimodal data during the course driving, The Proceedings of the Transportation and Logistics19 (2010), 283–286.

Umemoto

Liu

Narita

Kato

and Moriyama

, Steer-by-wire system for ultra-compact electric vehicle-fundamental consideration on the burden of the push side upper arm using EMG of anterior deltoid, Journal of the Japan Society of Applied Electromagnetics and Mechanics25(2) (2017), 155–161.

Liu

Umemoto

Kato

and Narita

, Muscle burden effect of ultra-compact electric vehicles equipped with steer-by-wire driving assist system, International Journal of Applied Electromagnetics and Mechanics59(3) (2019), 1143–1151.

Liu

Uchino

Ikeda

Endo

Peeie

M.H.B.

Narita

and Kato

, Driving assist system for ultra-compact EVs—fundamental consideration of muscle burden owing to differences in the drivers’ physiques, Actuators7(3) (2018), 44.

Abe

, Automotive Vehicle Dynamics Theory and Applications, Tokyo Denki University Press, 2008.

Izutani

and Shimizu

, Development of EPS energy consumption method, JTEKT Engineering Journal1007 (2009), 44–47.

Kizuka

Masuda

Kiryu

and Sadoyama

, Practical Usage of Surface Electromyogram, Tokyo Denki University Press, 2006.

Asai

and Tagawa

, Alteration in muscle fatigue caused by load-type differences during isometric flexion of human fingers, Transactions of The Japan Society of Mechanical Engineers Series C78(790) (2012), 2152–2161.

Chihara

and Seo

, Fundamental study on ergonomic design method based on comprehensive evaluation of muscle load using surface electromyography, Transactions of The Japan Society of Mechanical Engineers Series C77(776) (2011), 1477–1490.