Week 21 - November 21-27

The only tasks this week were concerned with testing and troubleshooting the Red Rover. At the start of testing we checked the voltage level of the battery and set a timer so that we would know how long the Rover can be operated before needing to be recharged. We started by testing the manual control of the Rover. We first checked to make sure that the phone application was communicating instructions to the Primary controller without the cart moving. Video 20.1 shows the Arduino IDE responding in real time to user input through the app.

Video 20.1

When we were confidant that the app was communicating with the primary microcontroller we connected the microcontroller to the rest of the system to test physical movement. Originally we tried using a 50lb load but the Rover was not able to handle that load. Then we reduced the load to 25lbs and were able to use manual operation to move the Rover wherever we wanted to. Video 20.2 shows the Rover being controlled manually through the phone application.

Video 20.2

Once we were confident that the Rover could be controlled manually, we started to work on testing object avoidance. We ran into a couple of road blocks (no pun intended) when the first couple of tries, the Rover ran into various objects. For testing we used empty cardboard boxes as our obstacles so that nothing would get damaged. Sometimes we even used ourselves to simulate what might happen if a person were to step in the way of the Rover. Videos 20.3-5 show some of the collisions that we experienced during early testing.

Video 20.3

Video 20.4

Video 20.5

After testing multiple circumstances we realized that the Rover was 'seeing' the objects, and did eventually respond and stop. However, the delay between seeing the object and actually stopping, unfortunately cause collisions. From these tests we went into the code to see if we could reduce that delay and allow the Rover to respond faster to sensor inputs. After retesting, we saw that we were successful and the Rover is now responding very well to obstacles and is able to correct its course to prevent collisions. Videos 20.6-20.8 show how the Rover stops and/or moves to prevent collision.

Video 20.6

Video 20.7

Video 20.8

During the filming of 20.8 the team was actually confused as to why the Rover stopped like it did. We were testing the Rover by having it reverse when it stopped unexpectedly. That is when we realized that the Rover had sensed the shelf behind it and stopped to avoid hitting it. It was one of those instances of happily discovering that something worked when you weren't even testing it.

The last mode that we tested was autonomous tracking using 'blind' and line of sight tracking. Videos 20.9-10 show the Rover responding to a call command from the phone app and moving towards Jesus (Holding the phone and User module). This function required surprisingly few adjustments. The only real issue we had was, once again, the reaction time between when the Rover gets within range of the user and when it actually stops. We fixed this by increasing the frequency of when the Rover checks the RSSI value so that it can sooner detect the -65dBm that lets the Rover know when it has reached the User.

Video 20.9

Video 20.10

As you can see from the previous two videos, the Rover is capable of moving towards the user autonomously, turning corners if required to reach the user from across a room. The only ongoing issue we experienced throughout our testing was the fact that due to the holiday week, we did not have access to a room with the carpet that the Rover was designed to move on. The slicker surface provided by the hard wood floors upon which we tested, created an issue with the mecanum wheels slipping during movement. This caused the motion of the Rover to be less smooth than we would have liked. We were able to test briefly on a small area rug to verify that tight packed carpets are indeed ideal for the Rover. However, the rug was not large enough to complete any other type of testing. We simply recommend that the user adhere to our parameters for use and don't try to operate the Rover on any slick surfaces.

Our testing this week was overall very successful and we are pleased with the results we received. After about 2.5hrs of continuous use, the Rover started to have reduced function and so we recommend that the user move the Rover to a wall to be charged using the on-board battery charger after 2.5 hrs of use. The actual battery life exceeds the system requirements by 1.5 hrs.