Determine if car is moving forward or backward

Question

I would like to know if there is someway i can figure out if the car wheel is rotating in clockwise or anticlock wise direction. Im working on a project which determines if car rolled back on a hill before the user accelerates forward. I tried this using gps co=ordinates to determine if there is roll back. As this roll back could be in cms before user moves forward, im not getting reliable results. The only way i can think of now is to depend on wheel rotation. if the wheel rotates backward on slope i would know that it did roll back a little. So is there a way for me to determine if wheel rotated in clock wise or anti-clock wise direction through CAN bus interface or something. Any help here would be greatly appreciated.

Answer 1

A quick search for various terms based on "how to read wheel sensors for direction," and "how to read wheel sensors" and a few other variations resulted in a mish-mash of answers.

There are devices available that connect to wheel sensors and output a canbus data stream. The manual for the linked device references that the device can be connected parallel with an existing ABS on the vehicle. The manual also contains a link to free download software to read the device on a computer. I was unable to determine if direction is part of that stream, but I suspect that it is, based on the results of another search:

Magneto Resistive (Active Wheel Speed Sensors)

A magneto resistive sensor has a sensing element that contains an output module that is mounted on a substrate and magnetic material. These magnetic rings are often located on the seal of the bearing. The magneto resistive sensor’s main advantage is the ability to sense direction and operate with a wide air gap. The sensor is more stable in high-vibration conditions and operates more efficiently than the reluctance sensor.

The wheel speed sensors can detect the direction of rotation of the reluctor ring and zero speed. The sensor’s compact size allows it to be incorporated into a hub bearing utilizing an alternating magnetic pole reluctor ring that is built into the rubber seal. A magneto resistive wheel speed sensor power input can range from 12V-5V from the ABS controller.

The sensor produces a square wave signal with a high signal of 1.93 volts, and a low signal of 0.64 volts. The high signal must cross 1.29V and the low signal must cross 0.97V to low. The main advantage is that the amplitude of the signal does not change.

Additional information suggests that one can use an oscilloscope to read the signal from the wheel sensor, the result of which would be a square wave. I'd expect that the spacing of the square wave may determine the direction of rotation but that's barely-founded conjecture. It's also possible that a pair of square waves are generated by the sensor and the difference between the two determine direction.

Another link I pursued more or less confirms my conjecture, by stating that the newer devices contain two sensors offset from each other:

The sensor assembly consists of two sensing elements mounted side-by-side with an amplifier chip built into the sensor assembly. The output from each sensing element rises and falls the same way as with a magnetic pickup, and the amplifier converts that to digital high-low voltage signals. Also, as before, the high-low switching frequency is proportional to wheel speed, and because the signal is always present, zero wheel speed can be detected. But because the sensing elements are next to each other, the two voltage signals are always slightly out of phase: one rises or falls just a few degrees after the other. That's how the sensor detects the direction of rotation. If the signal from element A lags behind the signal from element B, the wheel is turning clockwise. If B lags behind A, the wheel is turning counterclockwise. The exact same out-of-phase dual-sensor technique is used in radio knobs.