Project
| # | Title | Team Members | TA | Documents | Sponsor |
|---|---|---|---|---|---|
| 78 | Pitched Project: CfA Flying Area Accuracy Determination |
Alex Hu Bella Altenbach Juliana Temple |
Sarath Saroj | design_document2.pdf final_paper1.pdf other1.MOV photo1.jpg presentation1.pptx proposal2.pdf |
|
| Team members: - Alex Hu (alexxh2) - Juliana Temple (jtemple4) - Bella Altenbach (ialten2) # PROBLEM: The challenge faced right now is that the Intelligent Robotics Lab Facility would like to design a software that will analyze how consistently the position of a drone is able to be tracked throughout the Flying Arena based on the configuration of the motion tracking setup. The current motion tracker system (Vicon Tracker 3) gathers up to mm position accuracy, but it may be less reliable in some areas where the configuration does not allow for optimal observation. The goal is to see how the accuracy changes when going higher, lower, and further away into the arena and away from the cameras. Ideally, throughout testing the calibration of the motion tracking configuration can be improved based on where we identify these areas of high and low efficiency to be. # SOLUTION OVERVIEW: For our solution we will have to be able to actively track location and configuration of a test object or a flying drone using infrared LEDs. In order to track accurately we will be creating a calibration device using infrared LEDs (active marker) rather than reflective balls (passive marker) because the IR LEDs will be triggered from the flash of a camera. This allows location to be measured more reliably at a further distance. Additionally we would like to set individual IP addresses for each Led on the PCB so we can individually identify each marker and see the overall orientation during real time. A reference deck framework has already been made that has initial LED placements on all 4 arms for the main directions of the propellers, and we aim to design ours in a similar way. Using Vicon tracker 3 Motion Capture System and cameras, data on the location will be continuously recorded, and compared with the actual location of the calibration device. We plan on designing a software to measure the relative error in these measurements and assess where points of higher and lower accuracies are. Based on this, we will be able to reconfigure the camera locations in order to collect the most accurate position tracking data at all points throughout the arena. Example Reference deck: https://www.bitcraze.io/2019/09/the-active-marker-deck/ ## SOLUTION COMPONENTS: - PCB board - Infrared LEDS - Controllers to process data and orientation to actively see location - Vicon motion capture Camera for triggering LEDS and continuous recording - Vicon Tracker 3 software - Drone device/Test object to carry the PCB ## SUBSYSTEM 1: PCB and MARKERS - PCB board that contains multiple infrared LEDs that can be programmed into different configurations. These are active markers, and they activate upon camera flashes. Because LEDs emit light instead of just reflecting, they will be more effective in taking data throughout the whole volume of the arena. - Placing the LEDs in various configurations will help to get an accurate 3D location of the object, as they will allow the user to differentiate between up, down, left, and right. - This will be connected to the drone or test object discussed in Subsystem #2. - This will allow for testing of the calibration ## SUBSYSTEM 2: DRONE/TEST OBJECT - The drone can either be flown around the arena with the PCB attached, or the PCB itself can be carried throughout the arena to different locations. As our group is new to flying drones, the latter may be a safer option as to not harm any equipment. ## SUBSYSTEM 3: MOTION TRACKER and CAMERA SYSTEM - Vicon Tracker 3 Software -- This will allow the position of the object or drone specified in Subsystem #2 to be pinpointed - Vicon motion capture Camera -- The flash of this Camera will trigger the IR LEDS and continuously record data that can be analyzed by the user. ## SUBSYSTEM 4: SOFTWARE - Based on the actual location and the recorded location from what was recorded in Subsystem #3, we will design a software to determine how accurate the data is. - Based on this, we will analyze where the higher or lower areas of accuracy are and why. - Recorded data from the motion tracker setup will be streamed to the software over local WiFi. # CRITERION FOR SUCCESS: In order to reach success in this project we will separate the subsystems and take on each system individually to ensure that as we go each component functions as expected. First we will need to understand and improve our abilities using the Vicon software for the motion capture aspects, as none of us have prior experience using it. Next, we will need to design a PCB board that has IR LEDs strategically placed to gain the most efficient and readable directions from the test object moving around the arena. It should be programmable to test various configurations of the LEDs to achieve optimal calibration. Once this is done we will have to brainstorm and design a software program that will determine the accuracy of the motion-tracked locations versus the actual locations. Understanding the reason for why the accuracy is affected in certain places and creating a successful PCB board will be the strong points for our project to be successful. If time is permitted, repeated testing can be done to place the motion tracking cameras at different locations to improve the accuracy. |
|||||