Slippage Detection and Traction Control System May 10, 2004 Sponsors Dr. Edwin Odom U of I Mechanical Engineering Department Advisors Dr. Jim Frenzel Dr. Richard Wall Team Members Nick Carter Kellee Korpi Brian McConnel 0
Table of Contents ABSTRACT... II 1. PROJECT DESCRIPTION... 1 1.1 PROBLEM... 1 1.2 SOLUTION... 1 2. STATUS... 3 2.1 DESIGNED AND WORKING... 3 2.2 DESIGNED AND NOT WORKING... 3 2.3 DESIGNED AND NOT TESTED... 3 2.4 NOT DESIGNED... 4 3. METHOD OF SOLUTION... 4 3.1 TECHNICAL DESCRIPTION... 4 3.2 THEORETICAL BASIS AND FUNDAMENTAL RELATIONSHIPS... 7 4. VALIDATION PROCEDURE... 7 4.1 TEST PLAN... 7 5. RESULTS...8 5.1 OPERATING PROCEDURES... 8 5.2 VALIDATION RESULTS... 8 5.3 COST ANALYSIS... 10 i
Abstract This paper discusses the progress on the traction control system being developed for the University of Idaho s formula one race team. The system will detect slippage between the wheels and cut power in the engine accordingly. The current achievement is a simple version of a system, only able to compare two signals. The control algorithm can input a signal from the user control and two signals representing the wheel velocities. It compares the signals correctly and creates an output when slippage is above the allowable value. The system is going to be modified to compare four wheels using external counters. ii
1. PROJECT DESCRIPTION 1.1 Problem Each year the Formula Society of Automotive Engineers (FSAE) holds a formula one race for students from different universities to compete against each other. The University of Idaho participates in this race and in an attempt to increase their race standings they would like to implement a traction control system for their 2005 vehicle. This system must comply with the FSAE rulebook and the client, Dr. Edwin Odom, has specified additional constraints. Dr. Odom would like the system to be lightweight, durable, and user friendly. He would also like to have an override switch in case of system malfunction. The main objective of this project is to design a system that will aid the driver in keeping total control of the vehicle. Over acceleration at critical points in the race can lead to wheels spinning which can result in an increased race time or a vehicle spinout. This system will compensate for any over acceleration that may occur during start up or around corners, which will aid the driver in maintaining the control. 1.2 Solution The traction control system will monitor and compare wheel velocities to determine if wheel slippage is occurring. It will also have a user controlled slip allowance switch. If the actual slippage exceeds the allowable slippage, the system will send a signal to the engine control unit (ECU) to cut the spark or retard the spark timing. This will lead to a decrease in wheel speed, which will allow the wheels to regain traction from the road. Figure 1 shows the basic information flow for the system. 1
Figure 1: Block diagram of the traction control system The system will measure the velocity of the wheels using inductive velocity sensors. These sensors detect metal targets that are within their sensing range. The target used for this system will be a toothed wheel mounted on the axle near the vehicle s wheel. As the target turns, the sensor creates a digital output resembling a square wave. This signal will be buffered by a MOSFET to ensure the signal is of TTL levels and will then be input into digital counters, which will count the pulses of the signal. This count gives a representation of each wheel s velocity that can be used in the microcontroller algorithm for comparisons. The system s user control input enables the driver to select an allowable amount of wheel slippage before the system will go into effect. It will have an override switch that will disable the system if there are any malfunctions The inputs from the counters and the user control will be fed into the microcontroller. The microcontroller will check the counters at regular intervals and calculate the percent difference of the counts from each wheel. This difference will be compared to the user control input and if the allowable slip is exceeded then the microcontroller will send a high digital output to the ECU. 2
The ECU monitors and controls different engine parameters. The digital output from the microcontroller signals the ECU to cut the spark in the engine. It is also possible to use one of the ECU s analog inputs to retard the timing of the spark. Both options will decrease engine power, which will slow the vehicle s wheels. 2. STATUS 2.1 Designed and Working A basic version of the system has been designed and is working. The components that are prepared are the user control switch, sensor, MOSFET buffer, and the microcontroller. The microcontroller can take in the signals from the user control and sensors and determine if the signals percent difference is exceeding the user controlled allowable slip, but currently it only compares two wheel velocities. 2.2 Designed and Not Working A couple of digital counters are currently being tested. These components should be able to count the digital steps from the velocity sensors and output a binary count of the pulses that can be sampled by the microcontroller and then reset. The microcontroller would be able to read these counters instead of using the two external interrupts for counting. This will increase code efficiency by allowing the microcontroller to read four signals instead of two. The counters are being tested with a signal generator set to a square wave. Currently the counters are not producing a binary count. 2.3 Designed and Not Tested The output signal from the microcontroller has not yet been input into the ECU. The ECU can recognize a voltage level as a high or low digital input and a high digital signal can be programmed to cut the spark. A LED intended to signal the driver of measured slippage has also been purchased. It has 10 individual sections, each of which could signify a certain amount of slippage. Possibly 5% calculated slip per section. This will warn the driver if the system is about to take control or that slippage has begun. 3
2.4 Not Designed Eventually the control algorithm will be modified to check for slippage between four wheels instead of just two. A method of comparison is needed to be developed to take into account cornering. When a vehicle is traveling around a corner, the differential allows the outer wheel to move faster to makeup the extra distance. In the current algorithm, this would be viewed as slippage. Another method of cutting engine power is to adjust the spark timing. This can be accomplished by the ECU custom analog input. This input can provide a range of timing delays that would each correspond with a certain input voltage between 0 and +5V. The input voltage would have to come from the microcontroller. This range could be realized using the Pulse Width Modulators (PWM) as outputs from the microcontroller. 3. METHOD OF SOLUTION 3.1 Technical Description The user control switch is comprised of a rotary knob and a binary encoder. Four input pins are used on the encoder to create a two-bit output that interfaces directly with the microcontroller. The encoder runs on +5V to ensure a safe signal for the microcontroller. Each of the encoder s inputs is connected to a pull up resistor that ensures that the pins will remain high when not selected. The rotary switch connects one pin to ground at a time. Once a pin is grounded, the encoder outputs the binary value corresponding to the pin value. The use of two bits enables us to have four selections for slip allowance. The velocity sensors are of the inductive kind and sense metal targets with a radio frequency that is emitted from the coil away from the face of the sensor. As the metal targets enter the radiated field eddy currents flow in the metal. The oscillator requires energy to maintain the eddy currents in the targets, so as the targets approach the eddy currents cause a greater load on the oscillator. When the load becomes too great the trigger circuitry senses when the oscillator stops and changes the state of the switching device. The inductive sensor s circuitry can be viewed in Figure 2 as well as the process of acquiring the digital output signal. The repetitive nature of the toothed wheel results in 4
a square wave voltage output that can be counted by digital counters or the microcontroller. The main concern of the sensors is mounting them within their sensing range, which is specified before purchasing. Figure 2: Inductive velocity sensor and target Each tooth of the toothed wheel is a target for the sensor. This wheel was made of steel so the sensor could detect the targets. Some important factors that had to be considered when making the toothed wheel were the target (tooth) size and the distance between the targets. To get a clear output signal it was recommended that the sides of the targets should be equal to the diameter of the sensor, so in our test case the diameter of the sensor was 12mm, therefore the target size was 12mm. The recommended distance between targets was twice that of the target size, so our distance between targets was 24mm. To calculate the actual sensing range the nominal sensing range, which is the range specified in the catalog, had to be multiplied by a reduction factor. The reduction factor accounts for different target compositions. In our case the nominal sensing range was 4mm and the reduction factor for the stainless steel used was 0.85. This resulted in an actual sensing range of 3.4mm. 5
The square wave voltage output from the sensors is a fifth of a volt below their supply voltage of 13VDC, which is roughly equivalent to the vehicle s battery. Obviously, this is not a TTL level voltage, so a buffering device that would bring the voltage down to an acceptable 5V was implemented. This was done by using an n- channel MOSFET with a 10kΩ pull-up resistor. The signal was applied to the gate of the FET and a 5VDC signal was applied to the drain above the pull-up resistor. The output was taken at the bottom of the pull-up resistor and the source was grounded. The output was still a square wave, but with an amplitude of 5V instead of 13V. We are using the RMC 3110 Rabbit Microprossesor Core Modual that is centered around the Rabbit 3000 microprossesor. THe RMC 3100 is a very compact usnit and has many versital features in a very small area, making this a powerful addition to any system. The RCM3110 has a lot of IO grouped into 6 serial ports giving a total of 54 IO pins 46 of them are configurable, 4 are inputs only and the other 4 are outputs only. All of these IO pins are 5 V-tolerant making interfacing with other devices much easyer. The prossesor opperates at 29.4 MHz and had built features that help to lower EMI output of the circuit. The core module comes loaded with 256K of flash memory 128K of SRAM, quadriture encoder inputs, PWM outputs, pluse capture and measurement, serial ports and many other features IO features. Currently the microcontroller is set up to detect positive and negative transitions and every time there is a transition a count is incremented. With only the two signals beinging monitored. Once this ifrmation is collected over a 100 ms period the values of the counters are then read and put into data buffers. Once this is compleated the counters as reset and they begine to increment once more. From this the percent differnce is calculated and then used to compare to the defined percentage. We also have a blinking light that is incorperated to blink with an interval of 1/2 a second. The purpose of this light is to show that the microcontroller is doing something and that the program has not gotten stuck at some point in the program. This program is a very simple consept that shows the capablity of the microcontroller. 6
3.2 Theoretical Basis and Fundamental Relationships The user control switch currently has four allowable slip positions: 5%, 10%, 15%, and 20%. These values were picked based on an existing traction control system that was researched. The system will have this variation for the convenience of the drivers. The higher allowable slippage is for the more experienced drivers, this means the driver still has control if there is only a slight amount of slippage is occurring. The lower levels of allowance are for less experienced drivers since the system will take control sooner in a slippage situation. The inductive sensor gives a digital voltage output. This signal is fed into the microcontroller and each of the pulses is counted using the external interrupts. The signals are sampled at regular intervals to count the pulses. These counts are used to calculate the velocity differences between wheels. Then the difference is compared to the user defined allowable slippage. If the difference exceeds that of the allowable, the algorithm sets the output pin high. The output from the microcontroller will be input into the ECU. The ECU will be programmed to recognize a high digital input as a notice to cut the spark. This spark cut will decrease the power in the engine. This drop in power will transfer to the wheels, which will slow down. Once the wheels decrease in speed they will be able to regain traction and the vehicle will be under control. 4. VALIDATION PROCEDURE 4.1 Test Plan The switch was first tested with an oscilloscope to view the output pin configuration. Once the switch was getting original sequences for each knob position, the switch was input directly into the microcontroller. The microcontroller has a text display that shows the current settings of the system. The predicted value of allowable slip was displayed on the screen at each position. 7
The control algorithm was tested with the actual sensor signal and a waveform from a signal generator. The signal generator was set to a square wave at an appropriate frequency. The signal generator s frequency was varied to represent a wheel slipping. The text display showed how many pulses were counted in the sampling period and the percent difference of the tires. If the percent difference was greater than the allowable slip a LED illuminated on the prototyping board. 5. RESULTS 5.1 Operating Procedures The user control can be changed at any point throughout the test procedure and the system still operates as expected. 5.2 Validation Results The following figures show the results from the different testing procedures. Figure 1is an oscilloscope print out showing the digital signal that the sensor outputs, which is signal 1, and the buffered signal, signal 2, that is actually put into the microcontroller. The voltage supplied to the sensor was 13VDC and signal 1 of Figure 1 indicates that there was a voltage drop across the sensor of 1.7V. The data sheet in the Appendix suggests that the drop should be less than 2V, which is confirmed here. Signal 2 of Figure 1 is the original signal buffered through the MOSFET. The amplitude of signal 2 is the 5 volts that was applied to the MOSFET drain, which is an appropriate 8
TTL voltage. It was noticed that signal 2 was inverted, but this produces the same number of pulses, so this does not create any complications. Figure 1: Digital signal from sensor (1) and buffered signal (2) Figure 2 is an example of the information that is currently displayed on the computer monitor by the microcontroller, it shows the user defined allowable slip, sensor pulse counts, and the amount of slippage between the two signals. The percent control is the user defined allowable slip percentage, which is changed using a switch. Currently, the microcontroller detects only two signals and these signals represent the front and back left tires. The numbers 49 and 41 in Figure 2 represent the number of pulses counted over a 100ms interval. This case has an allowable slip percent input of 5% with the front tire spinning 8% faster than the back tire. In this case the microcontroller would send a signal to the ECU to cut the spark, but right now this action is represented by illuminating an LED. Figure 2: Microcontroller text output screen 9
5.3 Cost Analysis Microcontroller: $210 Miscellaneous (Budget $135) Rotary switch: $3.17 Variable LED: $3.17 Circuit boards: $1.80 10
Appendix A 0
1
2