Finding Petroleum - Using inertial navigation to track people

Join | Existing Member Login

You are Home » News » View Article

Using inertial navigation to track people

Monday, December 06, 2010 in Feature Articles

How well can inertial navigation – measuring acceleration and orientation – be used to track people in areas where they can’t get a GPS signal? We spoke to Tessella’s Dave Dungate

We’re all familiar with the way accelerometers and gyros inside Wiis and iPhones can be used to track our bodies. Could the same technology be used to track the position of people around ships and buildings, where there is no radio coverage?

If there is radio coverage then tracking people is easy – they just need a GPS receiver to work out where they are, and a way of sending the position data to a central base station.

But if they are inside platforms or ships, the GPS signal gets weaker and another technology is needed.

Tracking the location of people is very important – to know where they are in the event of an accident, and to measure where they have been (eg to track what they might have been exposed to).

Tessella has worked with the nuclear industry, which is keen to track individuals in buildings with thick walls. It was also asked by the defence industry to work out ways to track soldiers.

Inertial technology

Inertial technology means using gyroscopes and accelerometers, which can track how much the object is moving and in which direction.

Many of us are familiar with gyroscopes and accelerometers, because they are found inside our iPhones and Wiis.

The hardware – the accelerators and gyros – are available with low cost, low mass and low power, and do not need any infrastructure to work – just a means of recording data and downloading it.

In addition, the device needs a transmitter to send the data somewhere, a battery, and probably also memory, to store the data if it can’t be transmitted.

The challenge with the technology is that, with the current low cost hardware, the location measurement becomes inaccurate very quickly.

Tessella is running experiments, where it asks someone to wear a foot-mounted device with acceleromaters & gyros, walk around the office to the point they started at, and see where the computer thinks they are. For a 50m long office, “After a few minutes they are completely off the screen,” Dr Dungate says.

However with some computer processing, the accuracy can be greatly increased.

For example, the computer can detect when someone’s foot lands on the ground, and do some calibration around that. For example, the speed of the unit is known to be low when the foot is in contact with the ground.

By doing that, the accuracy is increased greatly.

A further processing step can be to use a map of the room. If the person is going from one room to another, he is probably going through a door, so the person’s position can be recalibrated.

If the system is showing that the person is making a movement which is physically impossible, such as walking through a wall, then that indicates a mistake somewhere.

With these steps, the accuracy of the person’s location can be reduced to under a metre, after 5 minutes of walking.

“We use advanced particle filtering techniques – to combine the information together,” he said.

For more accuracy, the inertial navigation data can also be used together with any other available data; for example, the person could also carry a GPS which receives a signal every time the person goes outside, and uses this to recalibrate the inertial navigation data.

In future, the system could also learn – about the movements the person makes. “That’s an advanced area that we haven’t looked at yet,” he said.

Such advanced techniques could increase performance for the more challenging problem of implementation on a mobile phone, which can be in someone’s pocket, hand or even being “played with”.. “People tip them up and sideways, leave them on the floor and pick them up again,” he said.

External Links
» Tessella