Effect of steering assistance control by external information feedback control and chassis control.
In this study, we report on the development of a steering assistance control system that feeds back information on the outside environment collected by laser sensors to the vehicle driver. The system consists of an emergency avoidance assistance control program that performs obstacle detection and avoidance, as well as a cornering assistance control program that operates by detecting the white lines painted on roadways. Driving simulator experiments were conducted in order to confirm the effectiveness of these functions, as well as to improve understanding of the synergistic effects of the steering assistance and chassis control functions: camber angle control and derivative steering assistance (DSA) control. In our emergency avoidance experiments, which were assisted by the obstacle detection function, the automatic override function successfully intervened to prevent accidents in situations where it determined that manual steering by the driver would be too late to avoid the detected obstacle. In our experiments involving cornering assistance by white line detection, smoother steering around curves was facilitated by the system's ability to set up an optimal approach earlier than could be expected by the curve recognition processes used by human drivers. In addition, vehicle stability at the approach and exit of a curve was enhanced by our steering system, which helps stabilize driving in critical cornering areas by effective use of camber angle control. Taken together, these steering assistance controls not only improve safety but also enhance the vehicle speed range because steering delays are minimized by the phase advances to actual steering behavior provided by the DSA control. It is expected that these systems have the potential to provide significant improvements in automotive safety.
CITATION: Yamaguchi, R. and Nozaki, H., "Effect of Steering Assistance Control by External Information Feedback Control and Chassis Control," SAE Int. J. Commer. Veh. 9(2):2016, doi:10.4271/2016-01-8104.
In recent years, external recognition sensors installed to enhance safety have been widely applied to automobiles, and it is now possible to reduce traffic accidents by incorporating controls activated by feeding back sensor information unobtrusively during manual driving operations. Thus, it is clear we are gradually approaching the driverless automobile society currently under widespread discussion.
However, the joy of driving and the genuine thrill of owning a car might fade away once driverless automobiles become commonplace. Additionally, determining responsibility in the event of an accident has become a major issue. Therefore, in this study, we report on the development of a steering assistance system that can reduce human errors during collision avoidance and cornering operations by providing the driver with information obtained via external recognition sensors.
It is expected that during manual vehicle operations, such steering assistance systems would provide the driver with external world recognition sensor information on obstacles and curves faster than could be provided by humans senses, thereby giving him or her some leeway to handle countermeasure operations. The purpose of this study is to confirm of superiority of a fusion of manual operation and assistance control.
Furthermore, powered motorization and automatic control of various automotive features is proceeding at an accelerating rate. Once complete motorization is achieved, vehicle motion control technology can be expected to evolve in wholly new ways. For example, it is now possible to control the tire camber angles by using electromagnetic actuators. Kobayashi et al. reported that as the lateral force characteristic with respect to the camber angle increases, the maximum value of the lateral force can be achieved by inclining the tire camber angle in the negative direction at the critical cornering area of the larger tire slip angle. Thus, by maintaining a large camber angle in critical cornering areas, performance and safety are improved.
Most electric vehicle (EV) steering systems are "steer-by-wire", which means they are motor-based systems that operate on electric signals instead of mechanical couplings. Research into such steering systems, which included formulation of the derivative term, was pioneered by Hirao,[2-3] who also reported that yaw response rates and lateral acceleration in a normal driving area could be increased by the use of such systems.
On the other hand, in a study by Nozaki that examined driver models at the time of drift area counter-steering, it was revealed that the drift cornering counter-steering delay could be improved by derivative steering assistance (DSA) control. Furthermore, they concluded that DSA control could be performed unobtrusively so as to cause no discomfort to the driver.
In the present study, the authors also explore a steering assistance control system based on external information feedback that is designed to expand safe driving limits by combining camber angle and DSA control.
Driving Simulator to Simulate Drift Cornering
In this study, experiments were conducted using a driving simulator coupled with a motion device that is capable of replicating extensive yawing behavior and lateral acceleration through triaxial (yaw, roll, and lateral translational) control, thereby allowing dynamic behavior such as spin or drift in critical cornering areas. The driving simulator consists of a personal computer (PC) for control, a motion device, and control panels. Triaxial control is performed based on the yaw rates and lateral acceleration generated via the CarSim full-vehicle driving simulator developed by the Mechanical Simulation Corporation (MSC) of the USA.
The driving simulator configuration is shown in Figure 1, while the motion device dimensions and performance are shown in Table 1. Black curtains were drawn around the test subjects to enhance the sensation of actually driving, and test subjects operated the simulated vehicle based on moving images of actual driving situations projected on the PC monitor.
Table 2 lists the main vehicle components and their degrees of freedom, while Table 3 shows the vehicle model parameters used in our experiments. The CarSim vehicle model considered herein has a front engine mount and rear drive (FR) layout that easily enters the drift state. For tire cornering force characteristics, we used the general tire characteristics shown in Figure 2. The maximum cornering force occurs at a slip angle of about 10[degrees], and the acceleration and deceleration characteristics are as shown in Figure 3.
When added simultaneously, the slip angle and ratio are calculated using the friction circle concept based on CarSim's combined characteristics calculations. This concept produces simulated driving behaviors that match actual vehicle conditions when limitations are exceeded. Thus, when accelerating during the critical cornering limits, the rear wheels' maximum lateral force decreases as the rear-wheel drive force increases and the vehicle, which is now unable to withstand the centrifugal force, begins to spin out of control.
However, skillful control of the direction of travel of the front part of the vehicle body through counter-steering can put the vehicle into a drift-cornering maneuver. The final behavior experiment in this study demonstrates a rear-wheel skid. In this paper, the grip limit is defined as the point where the rear wheels generate the maximum cornering force.
OUTLINE OF EACH CONTROL
Emergency Avoidance Assistance Control by Obstacle Detection
When an obstacle is detected by the sensors, the distance between the vehicle and the obstacle, L, and the angle of the obstacle, [theta], is as shown in Figure 4. The control provides assistance by feeding back the distance and angle of the detected obstacle. Active assistance is provided when the specified distance to the obstacle is less than the distance L. Left and right discrimination of the obstacle is carried out at an angle of [theta]. A block diagram overview of the process is shown in Figure 5. In this figure, steering assistance to the right is performed when there is an obstacle to the left, and steering assistance to the left is performed when there is an obstacle to the right. The proportion of manual steering and assisted steering was set at seven to three, which was selected in order to reduce the sense of discomfort imparted on the driver by steering assistance.
Cornering Assistance Control by White Line Detection
When a white line is detected by the sensor, the angle between the white line and the vehicle was set to "+[theta]" or "-[theta]", as shown in Figure 6. The control provides assistance by feeding back the distance and angle of the white line, as shown the block diagram in Figure 7. Steering assistance conditions at the time of a right turn are met when the angle -[theta] of the sensors observing the white line on the left is greater than the straight driving value. Conditions of the steering assistance at the time of a left turn are met when the angle +[theta] of sensors observing the white line on the right is larger than the straight driving value. The ratio between manual and assisted steering was set to half the steering input value.
Camber Angle Control
In camber angle control, the steering angle proportion of the negative camber angle where the tire is inclined in the direction of the turn center is controlled as shown in Figure 8. The critical cornering performance is improved by the effect of the camber thrust. The side force characteristic (value calculated by the 'magic-formula') in the negative camber angle is shown in Figure 9. In the driving simulation used in the present study, a negative camber angle of approximately 20[degrees] (corresponding to 1.2 times the cornering force) is added to the rear wheel when the body slip angle value exceeds the maximum cornering force. The camber angle control flowchart is shown in Figure 10.
Derivative Steering Assistance Control
As shown in Equation (1) when DSA control is applied, the phase of the front-wheel steering angle advances in proportion to the steering wheel angle velocity, when compared with the value that result from the driver's steering action. This improves the ease of steering and facilitates maneuvering stability improvements.
[[delat].sub.f] = [[delat].sub.H]/N + P * [[delat].sub.H] (1)
([[delat].sub.f]: Front-wheel steer angle; [[delat].sub.H] : Hand wheel angle; [[delat].sub.H]: Hand wheel angular velocity; N: Steering gear ratio; P: DSA constant)
Drift cornering and grip cornering are distinguished by the body slip angle in the driving simulator. Additionally, during drift cornering that exceeds a body slip angle of 10[degrees], the control that performs DSA (assistance constant P = 0.07) is added. The DSA control function block chart is shown in Figure 11.
The test subjects are listed in Table 4. The experiments were conducted with a total of four subjects: one subject occasionally drove (Subject A), two subjects were relatively familiar with automobile driving (Subjects B and D), and one subject had a driving license but hardly ever drove a car (Subject C). The subjects were asked to complete a 30 minute training session before the experiments in order to familiarize themselves with operating the driving simulator.
Emergency Avoidance Assistance Control by Obstacle Detection
In the emergency avoidance assistance program experiment, obstacle detection was carried out based on the situation shown in Figure. 12. The first driving experiment was carried out at a vehicle velocity of 80 km/h with a distance of 25 m set between Obstacles 1 and 2, during which a comparison of the steering assistance function and the manual steering was performed. Next, the same experiment was performed but with a comparison between the assisted steering function alone and the assisted steering function combined with camber angle control.
In the next stage, in order to perform the drift area experiments, the vehicle velocity was set at 120 km/h and 45 m was set between Obstacles 1 and 2. This experiment compared the assisted steering function combined with camber angle control, and the assisted steering function combined with DSA. Note that L was set to 20 m in all of the experiments. The experimental patterns are shown in Table 3. The experimental results shows data for Subject B, but similar tendencies were noted for all test subjects.
In the manual steering mode experiment, a collision with Obstacle 2 can be seen from the course trace graph of Figure 13. However, it can also be seen that when assisted steering function mode was active, Obstacle 2 was avoided smoothly.
From the yaw rate graph in Figure 14, it can be seen that faster avoidance response was facilitated by the early obstacle detection provided by the assisted steering function. Therefore, it can be concluded that the convergence achieved a stable result.
The effectiveness of the assisted steering function is extracted from the traveling locus graph shown in Figure 15. In cases where the assisted steering function was combined with camber control, participants were able to travel the course in a stable manner. Additionally, based on the yaw rate graph in Figure 16, combining the assisted steering function with camber control improved the critical cornering due to the addition of the camber thrust effect. As a result, the increase in the yaw rate after Obstacle 2 avoidance was suppressed, and a swift convergence resulted.
Looking at the traveling locus graph in Figure 17, it is clear that the obstacle that could not avoided when the assisted steering function combined with camber control was utilized, could be avoided when the assisted steering function combined with DSA was used. From the yaw rate graph in Figure 18, it can be seen that the yaw rate decreased faster because of the improved steering response made possible by the DSA function. Additionally, the improved counter steering is the result of the higher yaw rate, so an improvement in the vehicle speed range was possible.
Cornering Assistance Control by White Line Detection
In the white line detection cornering assistance program experiment, manual steering and assisted steering function comparisons were made at a vehicle velocity of 60 km/h in the R = 40 m course shown in Figure. 19. Next, in order to evaluate critical cornering area expansion, the course where R = 75 m, and a vehicle velocity of 86 km/h, is compared for assisted steering function and the assisted steering function combined with camber control.
Furthermore, the experimental patterns that resulted at 88 km/h when the assisted steering function combined with camber control, and the assisted steering functions combined with DSA were compared are shown in Table 4. Note that while the experimental results show data for Subject A, similar tendencies were observed for the other subjects.
As can be seen in the travel locus graph in Figure 21, manual steering and the assisted steering function show similar trajectories. Figure 21 also shows a yaw rate graph where it can be seen that the assisted steering function suppresses the steering wheel angle at the time of the corner entry for what is recognized as a fast curve. The convergence and stable yaw rate in the corner results in a faster exit.
Figure 22 shows a graph of the running locus. While the assisted steering function effect is infated in this course, it appears to follow the center of the same course produced by the assisted steering function combined with camber control. Figure 23 shows a yaw rate graph. Here, the yaw rate increase is suppressed when the assisted steering function is combined with camber control due to the critical cornering improvement provided by the camber thrust effect, and convergence is stable.
The running trajectory graph shown in Figure 24 indicates that the vehicle veered off the set course under the assisted steering function combined with camber control, thereby indicating an unstable course return. In contrast, it was possible to follow the course using the assisted steering function combined with DSA. Figure 25 shows a yaw rate graph. Although the camber angle control yaw rate increase is suppressed during cornering, convergence overshoots result because critical cornering exceeds the set limits. Better phase response resulted in faster steering due to the improved actual front wheel steering angle in the DSA, which results in faster yaw rate convergence and an improved travel area.
When vehicle sensors detect an obstacle, the emergency avoidance assistance provided by the obstacle detection function can react faster than human beings. Therefore, when operating the vehicle manually, the function makes it possible to provide unobtrusive cues that avoid the unsettling effects that might occur when the active steering assistance function overrides the driver, which may occur in situations where the system deems there is insufficient time to avoid the detected obstacle.
When the cornering assistance control by white line detection function is used, curve detection by the sensor prompts a smoother entry into the corner and facilitates a more stable exit from the curve. These steering assistance functions offer improved response times reductions that are below what are possible by human drivers. As a result, improvements to the vehicle's steering stability could be confirmed.
Camber angle control with steering assist reduced yaw rates by permitting appropriate variation of the camber thrust effect. Critical cornering and stable operation became possible near the limit region due to camber angle control.
The combination of DSA control with the assisted steering function resulted in improved responsiveness during cornering by advancing the phase of the front wheel actual steering angle against the steering input. This was made possible by the addition of a derivative item in the steering system.
Furthermore, counter-steering response in the drift region has been improved, and vehicle velocity limitations have been relaxed because it is now possible to make corrections to the return of straight line driving at higher speeds.
[1.] Kobayashi, H., Oyama, K. and Kanashima, M., "Measurement Technology of Tire Contact Patch under Actual Vehicle Driving Conditions," Transactions of the Society of Automotive Engineers of Japan Vol. 39, No. 6, Pages 35-40, 2008. (in Japanese)
[2.] Hirao, O., "On the Steering System which has a Derivative term for the Improvement of the Stability of Anthro-mobile" Journal of Society of Automotive Engineers of Japan (JSAE), Vol. 23, No. 1, Pages 48-54, 1969. (in Japanese)
[3.] Hirao, O., "Improved Dynamic Characteristics of Automobile Steering System. - Case of an automobile driven by a man -," Journal of Society of Automotive Engineers of Japan (JSAE) , Vol. 20, No. 11, Pages 995-1002, 1966. (in Japanese)
4. Nozaki, H., "Effect of Differential Steering Assist on Drift Running Performance," SAE Technical Paper 2005-01-3472, 2005, doi:10.4271/2005-01-3472.
Ryo Yamaguchi and Hiromichi Nozaki
Department of Mechanical Systems Engineering, Kogakuin University
2665-1 Nakano-cho, Hachioji-shi, Tokyo 192-0015, Japan
Table 1. Driving simulator parameters and performance Item Performance Actuator AC servo motor Control method Three-axis for roll, yaw, and lateral movements (feedback control by potentiometer) Main specifications Shake & rotation frequency: 0-3 Hz Roll motion Max. angle: [+ or -]20 deg Max. angular velocity: [+ or -]50 deg/sec Yaw motion Max. angle: [+ or -]90 deg Max. angular velocity: [+ or -]40 deg/sec Lateral movement Max. displacement: [+ or -]200 mm Max. velocity: [+ or -]240 mm/sec Lateral acceleration: [+ or -]0.7 G Size of device Width: 1525 mm, length: 2037 mm, height: 1800 mm Weight of device 400 kg Power supply Single-phase AC 200 V and single-phase AC 100 V Table 2. Vehicle model components and degrees of freedom Bodies Sprung mass body 1 Unsprung mass bodies (wheel carriers) 4 Rotating wheels 4 Engine crankshaft 1 Total 10 Degrees of freedom Sprung body translation 3 Sprung body rotation (yaw, pitch, roll) 3 Suspension stroke 4 Wheel spin 4 Powertrain (engine crankshaft) 1 Tire delayed slip (lateral, longitudinal) 8 Brake fluid pressure 4 Total 27 Table 3. Calculation parameters Width of vehicle (mm) 1795 Wheelbase (mm) 2780 Distance from center of front axle to center of gravity (mm) 1110 Height from ground to center of gravity (mm) 520 Vehicle mass (kg) 1370 Roll moment of inertia (kg[??][m.sup.2]) 671.3 Pitch moment of inertia (kg[??][m.sup.2]) 1972.8 Yaw moment of inertia (kg[??][m.sup.2]) 2315.3 Table 4. Subject characteristics Subject A B C D Driving More than More than More than More than experience 5 years 6 years 4 years 4 years Driving freqency Sometimes Everyday Rarely Everyday Table 5. Experimental pattern Assist pattern Definition Manual steer Manual steering 100% Steer assist Manual steering 70% Steering assistance 30% Steer assist + Manual steering 70% Camber control Steering assistance 30% Negative camber angle control (Vehicle slip angle over 6 deg) Steer assist + DSA Manual steering 70% Steering assistance 30% Derivative steering assistance (Vehicle slip angle over 10 deg) Table 6. Experimental pattern Assist pattern Definition Manual steer Manual steering 100% Assist steer Manual steering 50% Steering assistance 50% Assist steer + Manual steering 50% Camber angle control Steering assistance 50% Negative camber angle control (All the time) Assist steer + DSA Manual steering 50% Steering assistance 50% Derivative steering assistance (All the time)
|Printer friendly Cite/link Email Feedback|
|Author:||Yamaguchi, Ryo; Nozaki, Hiromichi|
|Publication:||SAE International Journal of Commercial Vehicles|
|Date:||Oct 1, 2016|
|Previous Article:||Piston cooling nozzle oil jet evaluation using CFD and a high speed camera.|
|Next Article:||Real world duty cycle development method for non-road mobile machinery (NRMM).|