MEMS: Room For Improvement

Last week while traveling in the Bay Area, I forgot my GPS so I used my iPhone 4s to guide me around. My rental car came with a Microsoft SYNC USB connection, so I thought I’d have no problems using my phone with the car system. I plugged in my phone, input my next destination and was on my way. Not quite.

MapQuest audio was not working. I had to unplug the phone from the car to hear the turn by turn navigation. The iPhone 4s uses an A-GPS or assisted GPS and has an integrated digital compass, providing additional directional information. But for some reason, the mapping application was continually re-calculating. I was on the route, but the navigation system thought I was somewhere else!

I drove to the ISQED SensorsCon Conference and Becky Oh, President and CEO of PNI Sensors provided some great insight on sensors and the real world. Nine-axis sensing is common today. The combination of a gyroscope, accelerometer and magnetic sensors provide the tracking data to determine up and down, side to side, and back and forth movements. Sensors can register how fast we’re moving, which direction we’re facing, and turning radius.

But sensors are far from infallible. How closely it follows our real movement and how quickly sensors can react to movement determine how accurate and responsive a system can be. Noise and outside disturbances can throw off the sensors’ ability to accurately collect data. The gyroscope, which measures the yaw, pitch and roll, and rotational rates, can suffer from gyro bias. The accelerometer which measures pitch and roll, gravity, velocity, and provides an absolute reference can, be affected by noise. A magnetometer measures yaw, heading and also provides an absolute reference but can be thrown off by magnetic disturbance. There can be a lot of magnetic disturbance in a car due to the metal of the vehicle and various sources of electromagnetic interference.

PNI Sensors specializes in magnetometers and algorithms for sensor fusion. The company has done a lot of testing and comparisons to improve the operation of magnetometers in sensor fusion devices. The two magnetic sensors available today are the Hall effect and Magneto-inductive magnetic sensors. The Hall effect “compass” sensor is the most widely deployed because it offers low power, low cost and a small size. However, these devices can be noisy and have low resolution. Magneto-inductive magnetic sensors are low power, low cost, and provide good resolution with less noise, but unfortunately it’s just not small enough.

Sensor fusion may not be perfect, yet. But companies like PNI are working on improving accuracy and performance. There are still a lot of other issues that need fixing. For example, when the map function was running, the iPhone doubled as a great hand warmer. Next will be improving the interoperability of the car’s infotainment system and the smartphone.

Joanne Itow, Managing Director