JP7569691B2 - Accelerator control system for autonomous vehicles - Google Patents

Description

The present invention relates to an accelerator control system for an autonomous vehicle that detects changes in the total vehicle weight, which directly affects the control response of the autonomous vehicle, by the accelerator response during autonomous driving without requiring any special equipment.

Acceleration control, which allows vehicles of different models and weights to travel in an orderly fashion in traffic, is a control that suppresses changes in the distance between the vehicle and the vehicle in front. To achieve this, an accelerator control system is required that can adapt to changes in vehicle weight and road gradient to provide the necessary acceleration. This is even more necessary when driving in a convoy with narrow inter-vehicle distances.

Automobiles are equipped with a driving performance diagram that shows the vehicle's power performance. Power performance refers to the driving performance that the vehicle exerts by overcoming the rolling resistance of the vehicle through the power of the engine, and this driving performance is shown on a driving performance diagram with vehicle speed (km/h) on the horizontal axis and driving force (kgf) and rolling resistance (kgf) on the vertical axis. This diagram is the basis for accelerator control.

Rolling resistance refers to all forces acting in the opposite direction to the direction of travel of a vehicle. Rolling resistance can be classified into the following four types based on the cause of its generation:
Air resistance Rolling resistance Gradient resistance Acceleration resistance

When driving at a constant speed on a level road surface, a and b act as rolling resistance. When driving uphill, c is added, and when accelerating, d is added. Therefore, when accelerating uphill, a, b, c, and d all come into play.

"Air resistance" (a) is the resistance force that increases in proportion to the square of the vehicle speed calculated by multiplying the frontal area of the vehicle body by the aerodynamic drag coefficient; "rolling resistance" (b) is the resistance force calculated by multiplying the total vehicle weight by the rolling resistance coefficient; "gradient resistance" (c) is the resistance force calculated by multiplying the total vehicle weight by the sine (sinθ) of the gradient (θ%); and "acceleration resistance" (d) is the resistance force that resists speed changes (acceleration).

In other words, the tractive force (F) generated by the engine, which depends on the throttle opening, is related to the sum of (a) air resistance related to the square of the vehicle speed, (b) rolling resistance equal to the total vehicle weight multiplied by the rolling resistance coefficient, (c) gradient resistance equal to the total vehicle weight multiplied by the sine (sin θ) component of the road gradient (θ), and (d) the inertial force equal to the vehicle mass calculated by dividing the total vehicle weight by the acceleration of gravity (9.81 m/ ^s2 ) multiplied by the current acceleration.

Non-patent document 1 is a textbook on automotive dynamics, and in the chapter on power performance it explains driving performance diagrams, but does not describe how to calculate the total vehicle weight or the required acceleration.

Non-Patent Document 2 reports on the identification and modeling of the longitudinal motion of large trucks, but does not describe how to calculate the total vehicle weight or the required acceleration.

Patent document 1 shows a three-dimensional calculation diagram for estimating vehicle weight and calculating accelerator opening, but it requires a process of tracing the diagram, and there is room for improvement.

Katsuzo Kageyama and Ichiro Kageyama, Automobile Dynamics, Riko Tosho, ISBN 4-8446-03536-6 Fujio Momiyama et al., “Identification and modeling of longitudinal motion of large trucks” Journal of Automotive Technology, Vol. 43, No. 2, March 2012, No. 20124209, pp. 211-216

JP 2020-11555 A

The above-mentioned prior art documents are unable to determine the change in total vehicle weight while driving. As a result, it is not possible to perform accurate accelerator control during automated driving, including platooning.

The accelerator control system of the autonomous vehicle according to the present invention is equipped with a calculation formula (diagram) that detects changes in the total vehicle weight, which affects the vehicle's longitudinal motion (power performance and braking performance) and lateral motion (handling stability), from the accelerator response. Vehicle motion is determined by the action of force, which generates acceleration, which determines speed and displacement (distance), so "controllability of acceleration is essential." The accelerator control system of the autonomous vehicle according to the present invention is also equipped with a calculation formula (diagram) for the accelerator opening, which determines the acceleration of the longitudinal motion.

Specifically, acceleration tests are conducted with an empty vehicle, with the accelerator opening kept constant at 25%, 50%, 75%, and 100%, and acceleration versus vehicle speed is measured to create a "vehicle speed-acceleration diagram" of vertical axis acceleration versus horizontal axis vehicle speed. Next, the transmission is placed in neutral from the high-speed range, the accelerator is released and the vehicle is coasted while decelerating, and deceleration versus vehicle speed is measured to create a "coasting-deceleration diagram" of vertical axis deceleration versus horizontal axis vehicle speed. The coasting-deceleration diagram is a quadratic curve of y= ^ax2 +b. The resulting "a" is the aerodynamic drag coefficient, and "b" is the rolling resistance.

If you change the sign of the coasting deceleration diagram from "negative" to "positive" and add it to the previous "acceleration-acceleration diagram" for accelerator openings of 25%, 50%, 75%, and 100% and overlay it, you will get an "all-speed diagram" of vertical axis acceleration versus horizontal axis vehicle speed, with accelerator openings of 25%, 50%, 75%, and 100% as indicators.

The obtained "total acceleration diagram" for 25%, 50%, 75%, and 100% plots will draw a hyperbola of "xy=a _xy " or "y=a _xy /x" where "x" is the vehicle speed on the horizontal axis and "y" is the acceleration on the vertical axis. Here, "a _xy " is a hyperbolic constant. If this "a _xy " is obtained depending on the accelerator opening "%" and set as "a _xy% ", it will become formula (6).

In the above formula (1), F, a, b, c, and d are all "forces" in units of (N) or (kgf), but if you consider them as "accelerations" instead of "forces", it is convenient because "it is possible to understand them without the need for special equipment." When considered in terms of acceleration, formula (1) is based on the vehicle weight (e.g., unladen weight) at the time of the preliminary experiment, so when expanded to a loaded state of n times the unladen weight, formula (1) becomes formula (7).

By rewriting equation (10) as equations (11) and (12), we can plot the graph shown in the center of Figure 5, which will be described later. From this graph, we can determine the accelerator opening relative to the required acceleration, and the load ratio (n) relative to an empty vehicle.

According to the present invention, changes in the total vehicle weight, which is a control parameter necessary for controlling the longitudinal and lateral movement of an autonomous vehicle, can be detected by the accelerator response during autonomous driving without requiring special equipment, and the accelerator opening angle required to achieve the target acceleration can be obtained.

FIG. 13 is an explanatory diagram of an example of a running performance diagram. FIG. 13 is an explanatory diagram of an example of acceleration resistance expressed as acceleration obtained from an acceleration experiment. FIG. 13 is an explanatory diagram of an example of air resistance and rolling resistance expressed in acceleration obtained from a coasting experiment. This is an explanatory diagram of an example in which the measured acceleration resistance and the measured coasting resistance were added together, the sum was verified to be a hyperbola, and the hyperbolic constant was obtained. FIG. 13 is an explanatory diagram of an example of a weight estimation formula and an accelerator control formula. FIG. 2 is an explanatory diagram of how to calculate gradient resistance. FIG. 13 is an explanatory diagram of a control flow.

The accelerator control method of the present invention, which estimates the change in the vehicle's total weight and generates a target acceleration, can also include a means for detecting the accelerator opening as an engine control input and the resulting longitudinal acceleration, as well as a means for detecting the vehicle speed as a result of the longitudinal acceleration, and a means for detecting the road gradient.

A method for detecting the target vehicle weight and an accelerator control method for outputting a target acceleration will be described below with reference to FIG.
FIG. 1 shows a driving performance diagram equipped in the control unit (ECU) of each vehicle. Here, an example of a large 4-axle truck with a GVW of 25 tons is used for explanation. The horizontal axis shows the vehicle speed, the left vertical axis shows the tractive force and running resistance in units of Newtons, and the right vertical axis shows the engine speed. I, II, III, IV...IX, X, XI, XII show the relationship between engine speed and vehicle speed in each gear. The mountain-shaped lines aligned from the upper left to the lower right of the diagram show the tractive force of each gear stage obtained by multiplying the engine torque curve by the gear ratio. The line that envelopes these tractive forces of each stage is a hyperbola (xy=a _xy ). The slightly upward-sloping curves of 0%, 3%, 5%,...40%, 45% show the road gradient.

For example, the tractive force with the accelerator fully open at a vehicle speed of 65 km/h can be read as 14,550 (N). Of this, the rolling resistance on a flat road with a 0% gradient is 2,000 (N), accounting for 14%. Because it is a flat road, this 14% is interpreted as rolling resistance (b above) and air resistance (c above). On a 3% gradient this becomes 59% of 8,640 (N). The remaining 41% is excess tractive force on a 3% gradient.

The magnitude of acceleration due to this excess traction force is calculated to be 0.229 (m/ ^s2 ). At 65km/h, it is possible to drive in 10th, 11th, and 12th gears on a 0% gradient, but it is impossible to drive in 12th gear on a 3% gradient, and neither 11th nor 12th gear is possible on a 5% gradient, so it is necessary to shift down to 10th gear. It is necessary to compile a gear shift plan for the route based on this reading, but a plan that can adapt to fluctuations in load, gradient, and vehicle speed at the time of operation is necessary for control. In order to make this possible without special equipment, "acceleration" is used instead of the "force" of traction force and running resistance. Acceleration can be detected from the wheel speed of the CAN or a gyro, and is normally equipped in self-driving vehicles.

Figure 2 shows examples of acceleration resistance obtained from actual vehicle acceleration experiments, in order to use "acceleration" instead of the "forces" of traction force and running resistance. From the top left, it shows 100%, 50%, and 40% accelerator for an empty vehicle, and from the bottom left, it shows 100% and 50% accelerator with a 12-ton load. The results show that the generated acceleration is proportional to the accelerator opening and inversely proportional to the total vehicle weight, as would be expected.

Figure 3 shows an example of air resistance and rolling resistance expressed in acceleration obtained from a coasting experiment. (A) is an empty vehicle, (B) is a vehicle with a load of 12 tons. Rolling resistance depends on the load, and air resistance proportional to the square of the vehicle speed is captured as per theory. In Figure 3, the line segment with large amplitude in the vertical direction represents the G meter acceleration, and the line segment with small amplitude in the center represents the vehicle speed differential acceleration.

Figure 4 is a diagram verifying that the sum of the actually measured acceleration resistance and the actually measured coasting resistance from Figure 3 forms a hyperbola. The horizontal axis is vehicle speed x (m/s) and the vertical axis is acceleration y (m/ ^s2 ), with the hyperbola 22 at 100% accelerator with an empty vehicle, 7.5 at 50% accelerator, and 40% accelerator. Furthermore, with a 12-ton load, the hyperbola is 12 at 100% accelerator and 3.5 at 40% accelerator, proving that the acceleration is proportional to the accelerator opening and inversely proportional to the load.

Figure 5 is an explanatory diagram of an example of the weight estimation formula and the accelerator control formula. It illustrates the flow of the expansion from formula (1) to formula (7), and from formula (7) to formula (13).

FIG. 6 is an explanatory diagram of how to calculate the gradient resistance. In order to determine the accelerator opening that produces the target acceleration or command acceleration while driving, it is necessary to detect the vehicle speed and road gradient at each location and respond to the detection of the vehicle's own weight. A three-axis accelerometer is provided. The reading of the longitudinal acceleration (G _x(static) ) when stationary on a horizontal road surface is -9.81xsineθ(m/ ^s2 ) in equation (14) when stationary on a slope with a gradient (θ). When driving on a slope, the forward driving acceleration is added to the static acceleration (G _static ) in equation (15).

The bottom left of Figure 6 shows a graph comparing the gyro pitch angle with the results of experimentally verifying equation (16) on a 4% slope. The four graphs, from the top, are vehicle speed, acceleration, road gradient, and gyro pitch angle. The acceleration graph shows the longitudinal acceleration (Gx) measured by the accelerometer in green, and the differential value of the wheel speed in red. A zero differential value of the wheel speed (red) indicates climbing at a constant speed. The third graph from the top (road gradient) shows the calculation results when the longitudinal acceleration and wheel speed differential values of this graph are substituted into equation (16). A gradient of 4° was detected. The fourth graph from the top shows the gyro pitch angle for comparative verification. Although it contains an initial value error, it can be seen that it is possible to detect the slope gradient using the gyro pitch angle as well.

Figure 7 is an explanatory diagram of the control flow. It has a flow that outputs the total vehicle weight, which changes with the load, by calculation, and a flow that outputs the accelerator opening degree corresponding to the required acceleration by using a nomogram. The upper part of the figure shows the flow of the calculation formula, and the lower part shows the flow of the nomogram.

The flow chart of the calculation formula in the upper part will be explained from left to right. The road gradient (θ) is detected by reading the wheel speed from the CAN and the longitudinal acceleration from the accelerometer and substituting them into equation (16). In parallel, the pitch angle (θ) from the GPS or gyro is taken in and selected for use. This θ, the wheel speed x (m/s) and the test acceleration are input into equation (13), the gradient resistance of c is updated by θ, the air resistance of a _d x ² and the hyperbolic constant of F, which depends on the vehicle speed, the test acceleration input is input into the acceleration resistance of d, the load ratio n is estimated, and the total vehicle weight is estimated. The total vehicle weight is substituted into equation (12) to determine the accelerator opening.

The lower part of the calculation diagram will be explained. This diagram is two-dimensional, with the accelerator opening (F) on the horizontal axis and the acceleration (d) on the vertical axis. For example, the vehicle speed range is determined as low speed range, medium speed range, and high speed range. The preliminary experiment data (thin line) conducted with an empty vehicle is described in a linear equation. The intersection point of this line and the vertical axis corresponds to "- a _d x ² -(b+c)" determined by air resistance, rolling resistance, and gradient resistance, and the slope of the line corresponds to the vehicle weight. In actual operation, when the vehicle is loaded, the accelerator opening and the generated acceleration are input at several points to obtain a correlation linear equation for those several points. The slope of the obtained line is the ratio (n) of the slope of the line to the slope of the empty vehicle linear equation. The total vehicle weight in the loaded state can be estimated by multiplying the empty vehicle weight by the obtained n. The accelerator opening that responds to the required acceleration (vertical axis) can also be obtained from the estimated linear equation.

From the above, it is possible to realize an accelerator control system that can adapt to changes in the vehicle's own weight and road gradient to achieve the required acceleration. The method described here is not limited to freight vehicles or passenger transport vehicles, and is a universal method that can be applied to all vehicles regardless of the type of power source.

Claims (1)

An accelerator control system for an autonomous vehicle that calculates the accelerator opening in response to a required acceleration, the accelerator control system being characterized in that the following are stored in an autonomous driving ECU: a calculation formula that detects the road gradient by reading the wheel speed and the longitudinal acceleration from an accelerometer; the following formula (13) that calculates the load ratio (n) from the traction force (F), air resistance coefficient (ad ₎ , vehicle speed (x), rolling resistance (b), gradient resistance (c) and acceleration resistance (d) generated by the engine which depend on the accelerator opening; and a calculation formula that detects the loaded total vehicle weight (W) from the calculated load ratio (n) and the unladen total vehicle weight, and calculates the accelerator opening from the detected loaded total vehicle weight and the required acceleration.

Here, a is air resistance, a _d is air resistance coefficient, b is rolling resistance, c is gradient resistance, d is acceleration resistance, x is vehicle speed, and F is a+b+c+d.

Images