Presentation on theme: "Dan Garber, Jacob Hindle, and Bradley Lan Advisor: Dr. Joseph Driscoll"— Presentation transcript:
1Dan Garber, Jacob Hindle, and Bradley Lan Advisor: Dr. Joseph Driscoll Autonomous Quadcopter with Human Tracking and Gesture Recognition (QUADHTGR)Dan Garber, Jacob Hindle, and Bradley LanAdvisor:Dr. Joseph Driscoll
2Outline Project Summary Previous Work Preliminary Work Quadcopter InvestigationEquipment and Parts ListBlock DiagramUltrasonic SensorGPSSchedule Overview
3Project Summary Human Tracking Gesture Recognition Live Video Feed & Data Readings via WifiObstacle Avoidance and Auto-StabilizationR/C Manual OverrideThe goal of this project is to create a quadcopter that can autonomously track and follow a particular human as well as respond to gesture-based commands. The human tracking will rely on sensor data such as video images and GPS coordinates. The quadcopter will also have a wifi connection, allowing live video feeds and data to be streamed over the internet. Autonomous flight will include auto-stabilization, and obstacle avoidance. As an added safety measure, there will be a radio control (R/C)-based manual override.
5Gesture Recognition BeagleBone Black I2C Communication Camera Bicycle Hand SignalIncrease Following DistanceDecrease Following DistanceLand
6Live Video Feed & Data Readings via Wifi Video Captured by CameraGPS Data
7Obstacle Avoidance and Auto-Stabilization Ultrasonic SensingImage ProcessingInertial Measurement Unit
8R/C Manual Override Required by FFA Turnigy Power System 9X 2.4G 9 Channel ControllerR/C MUX CircuitManual override- There shall be a manual override, in the form of a standard R/C system, to switch from autonomous control to manual R/C control. The Turnigy Power System 9X 2.4G 9 Channel controller will be the initial platform and will be upgraded as needed.
9Previous Work Junior Project Introduction to NMEA Messages GPS ReceiverATmega Based MicrocontrollerIntroduction to NMEA MessagesSerial Communication via UART for GPS Data
10Previous Work Junior Project Roboboat Senior Project Sharp Infrared Sensors for Edge DetectionATmega128 MicrocontrollerRoboboat Senior ProjectMoving Average Low-Pass Filter
11Preliminary Work Quadcopter Investigation Equipment and Parts List Block DiagramUltrasonic SensorGPS
12Quadcopter Investigation Two ESC FailuresLack of Information/Poor DocumentationBlack Box ControllersFC1212-S (Flight Controller)AHRS-S (IMU)
13Considered Alternative Options Option 1 - Repair X650CF, New XAircraft PartsOption 2 - Keep X650CF Frame, New Generic PartsOption 3 - Purchase a New QuadcopterAPM:CopterELEV-8 ParallaxParrot AR DroneDJI Phantom
18Ultrasonic Sensor Specifications XL-MaxSonar-EZ3 Ultrasonic SensorSize: 2.2cm x 2.0cm x 2.5cm (6.1g)Range: 20cm - 750cm (datasheet)Lab Test: Max Distance = About 500cmReading rate: 10 HzTravel Limitation w/ Six Ultrasonic Sensor = 8.3m/sTravel Limitation is calculated by taking 500cm max distance reading and dividing it by 0.6s for its refresh rate across six ultrasonic sensors in a low noise chaining configuration.
19Preliminary Work - Ultrasonic Sensor PWM Signal OutputLow Pass Filter of SignalAnalog Signal Output and ADCLow Noise Chaining of Multiple Sensors
20Low Noise Chaining Diagram Pin 5 on each sensor fires off a 100ms pulse when pin 1 is tied to ground.
21Adafruit GPS Specifications Sensitivity: -165dBmTypically around -160 or -165dBmUpdate Rate: 10Hz MaxStandard ≥ 1HzPosition Accuracy: 1.8mVelocity Accuracy: 0.1m/sTravel Limitation is calculated by taking 500cm max distance reading and dividing it by 0.6s for its refresh rate across six ultrasonic sensors in a low noise chaining configuration.
22Preliminary Work - Adafruit GPS Program for ATmega328PShift to BeagleBone BlackTravel Limitation is calculated by taking 500cm max distance reading and dividing it by 0.6s for its refresh rate across six ultrasonic sensors in a low noise chaining configuration.