This week was mostly spent on finalizing our requirements and design for the project for the upcoming report.

As a group we worked on setting up our various sensors with the raspberry pi and testing them out.

Based on testing results and the distance and angle specifications for the ultrasonic and lidar sensors, I came up with a new design and worked out all of the math for a new sensor approach with 3 Lidar sensors, that we’ve decided to go with. The new approach is to essentially have 3 Lidar sensors on top of each other, with one angled directly in front of it, one angled at 4.5 degrees to the right, and one angled 4.5 degrees to the left. Because each lidar sensor has a sensing angle of 2.3 degrees, and should be accurate at a distance of 10m, this approach allows us to sense a 2m range at 10m. There will be 2, 15-inch gaps in the line of sight (between the sensors views), but nothing we are sensing should be thinner than 15 inches, so that shouldn’t be a problem, even if a different sensor is picking the person as they approach (we just need the closest one). With this approach, we will be able to meet our 8-10m requirements, and we have settled that there should be a new requirement that there is at least 1m on each side of you between continuous barriers on either side of you (i.e. trees in a forest). This approach stemmed mostly from the results of testing the ultrasonic, which proved it was shaky over 6m, and the angle was a bit larger than advertised.

I didn’t get to spend as much time on object recognition work this week because there were more urgent parts of the design to spend time on, but I would like to work on that more next week.