VIDEO: Sensor-based system for tracking animal movements

Posted on May 10, 2017


Annie Hollis, Katie Fisher, Emily Santilli, and Athena Fung built a prototype sensor-based system for tracking animal movements for their capstone design project.

  • Video transcript:

    So can you tell me the name of your project and a bit more about it?

    So we're a sensor based system for tracking animal movements and we're working with the Psychology Department at Queen's and one of the tests that they use to monitor anxiety and anxiety treatments in animals is an elevated plus maze. So you can see right there an elevated plus maze is kind of a maze it's more of a cross and it has two closed arms and two open arms and what they'll do is put the rat down in the middle of the maze and they'll give it 5 minutes to wander around and see where it goes. So normally animals like rodents display a thigmotaxis bias so they kind of slink along the wall away to stay safe from predators but if an anti-anxiety treatment is working then they'll spend more time out in the open where they will feel less safe and more curious.

    So what they do after a test is they'll look at footage which will look somewhat like that and they'll go through and they'll score how long the rat spent in each different type of arm and how many entries it made.

    But the big problem is that the standard for making an arm entry is that it has to have all four feet cross the line out of the choice point in the middle. So it's kind of hard to tell and make it consistent between different researchers looking at the footage.

    And how do you find the link between the Electrical and Computer Engineering and to your other partner?

    All of us have interest outside of just our home discipline of Electrical and Computer Engineering. So we're hoping to find something to give us a bit of context for the theory that we're learning and we found that working with more Biomedical or Psychology application of our research is useful because it puts the system and the sensors and stuff into real world context and gives us the chance to actually apply them and see how it works with real users.

    So when you were going through the design of your project what were some of the considerations that your stakeholders had requested?

    Working with animal test subjects it was really important to consider how our system would affect their behaviour. So one of the things they said that is if there were loose wires on the system they would chew those wires or big lights or sound would distract them and alter their behaviour. So we wanted to make sure that there were no inconsistencies between the existing testing protocols and what our system would do. So one thing we really wanted to do was we were really considering a wearable unit like this.

    And we were concerned that maybe that would affect the animal's behaviour but we consulted with the Psychology department and they said that actually they habituate very well to harnesses like this.

    Another good thing about our harness is the electronic component is actually removable. The harness can be washed between test subjects and then also the foam coating we have on the maze can also be washed and cleaned.

    Can you tell me a bit more about the technical aspects?

    So the first sensor we used to implement our design was the force sensors which are in the middle of the maze which we call the choice point. When the rat is on the choice point the sensors will give us a 1 for "rat is on the choice point". And when it leaves and enters an arm it will give us a 0. So one of the points of the test is to count the amount of times it enters an arm and to count how much time it spends in the arm.

    We've heard a little bit about the overall scope of your design, but can you tell me about the technical aspects of it?

    So our system has two parts to it. There's four force sensors that are on the maze and we use those sensors to determine when it has actually entered an arm so when it leaves the sensor we can count an arm entry and we need to determine which one and that's why we have the wearbles component which is going to be on the rat and it has the IMU sensor which stands for Inertial Measurement Unit and it has an accelerometer, a gyroscope and a magnetometer on it. So we use that to determine where the rat is. Right now we can use this data to determine which arm it has entered in.

    What would you say are the next steps either for yourselves and your team and or for your stakeholders?

    So one thing that our stakeholders mentioned is that it would be nice to have a visual output along with the output that we have already and we are currently actually working on that using the data from the IMU we're trying to output exactly where the rat is in the maze and we've gotten results but so far there's a lot of noise that comes with the data including problems with the bluetooth. We still have to make the data more consistent and more reliable.

    We'd like to see how else we could apply our system. For example: right now it's only for rats, they also do perform elevated plus mazes on mice as well.

    And most of our requirements right now are outlined by the Psycology department here at Queen's but we could also talk to other departments and see what other types of behaviour that they're interested in cause we could get a lot of from the IMU.

    We can determine if its rearing against the wall or if it's just staying in one place or just hyperactive. So there's just a lot of potential.