One to ten. Modern IMUs

In the world where technology is seamlessly integrated into our daily lives, sensors play an indispensable role. They're in your smartphone, embedded in your car, integral to your smart home devices, and are the cornerstone of the burgeoning Internet of Things (IoT). Among the plethora of sensor types, inertial sensors hold a special place, especially for those of us fascinated by the intricate dance of physics and engineering. My own journey with MEMS (Micro-Electro-Mechanical Systems) accelerometers during my PhD has given me a deep appreciation for these tiny yet mighty devices.

Inertial sensors have evolved remarkably over time. They've progressed from simple 3-axis systems detecting linear or angular forces to sophisticated units capable of managing up to ten different functions. But what exactly entails a 'tenth degree of freedom' in these sensors? Let's delve into the captivating world of MEMS sensors to find out.

MEMS sensor development got a huge boost of capital due to the interest coming from the automotive industry. Airbags are critical components of automotive safety and a moment of impact is exactly what an inertial sensor can pick up and trigger the life saving device. In a time before MEMS, one common type was the "ball-and-tube" accelerometer. This device consisted of a small metal ball housed in a tube. Under normal conditions, the ball rested at one end of the tube. In the event of a sudden deceleration, such as during a collision, the inertia would cause the ball to move through the tube and make contact with an electrical switch at the opposite end. This contact would complete a circuit and trigger the airbag deployment.

Linear accelerometers appeared first. These devices consist of a proof mass and a set of electrodes, often in a comb pattern, against which a parameter like capacitance is measured against. These devices quickly reached a high level of sophistication.

Some of these devices, like the MPU-6050, can stay relevant on the market for more than a decade, some of the never ones are extremely compact i.e. Bosch BMA 6xx series and yet fundamentaly the same. These devices enable accurate measurements of linear forces, but are not good at angular measurements. This is where the MEMS gyro comes in. The Gyro is a MEMS inertial sensor that has two proof masses in a mirrored arrangement, which allows the sensing of the change of angle which, when combined with the linear acceleration data, can make navigation that much more accurate.

The weakness of the Gyro is that it tells you the change of the angle, but not the angle itself. This a bit of an issue, as small changes that are below threshold can very much build up to a complete rotation and the Gyro will be non the wiser! Well, good thing we have a frame of reference against which we can find our bearings. I am talking about the magnetic poles and MEMS magnetometers! The magnetometer is quite similar to the linear accelerometer but with a magnetised elements on the proof mass the will undergo a change in Lawrence force as their position relative to the magnetic poles changes.

The advantage of this sensor is that it reports the actual angle and is quite stable. But that makes a total of 9 degrees of freedom. What is the 10th? Well, it isn't an inertial measurement, but it certainly helps with navigation! It's pressure. Absolute pressure. 1 atmosphere at sea level, more below sea level and less and less as you go up. Typically a MEMS pressure sensor is a membrane over a hermetic cavity, as the ambient atmosphere pressure changes this causes a deflection of the membrane and if the bottom of the cavity and the membrane are electrodes this can act as a variable capacitor.

So have these sensors reached perfection? Well, no! While they probably won't change much in terms of what they measure, there is plenty of room to grow in terms of what the ASIC does. ASICs can be made to be very efficient in certain calculations. Such as matrix calculations in all sorts of Machine Learning! And that is a topic for another day.