Team 14 : Jonathan Pham, Pritesh Solanki, Po-Han Wu, Josh Schulte, Joseph Nguyen
Background Design Requirement Proposed Design Block Diagram I/O Specification Criteria
Design a system using multiple sensors that i nterface to a TI microprocessor. Use a Mobile Robot Platform Implement One Game & One Task Utilize Open Source Software
Multi Sensor System: Microsoft Kinect Processor: Beagle Board Mobile Robot: Roomba Game: Laser Tag Task: Intruder Detection Open Source Software: Microsoft SDK, TIesr (Speech Recognition)
Multi Sensor System: Microsoft Kinect Processor: Beagle Board Mobile Robot: Roomba Game: N/A Task: Intruder Detection Open Source Software: Open Kinect, Player TIesr (Speech Recognition) Modified Design
Beagle Board Kinect VoiceVideoInfrared Roomba
Subsystem Criteria Object Recognition Detect object of size 3 cm by 3 cm. At Distance of 5 feet Movement Moving at full speed: 250 mm/sec without hitting any object. Speech Recognition: Identify the key command 80% Aim and shoot: Detect target within 1 meter Within 5 sec aim at object location and shoot. Accuracy 90%
Chosen robotics platform: The Roomba –Smooth mobility; easy to navigate –Low-cost –UT lab provides programming capability –Provides stability for kinect (unlike robosapiens)
Chosen sensor array: Kinect –Integrated with all desired sensors –Unrivaled performance –No integrated microprocessor –Low-cost solution –Many open-source software implementations (MSFT)
Chosen microcontroller: Beagle Board –Powered by TI OMAP3530 ARM processor –Low-powered –Low-cost –720MHz ARM Cortex- A8 Core –256 MB LPDDR RAM memory –Contains many desired peripherals
Intruder Detection Using Player, beagleboard will control the movement of the roomba. Using Open Kinect, kinect will capture video, and beagleboard will process for intruder detection. If detected, beagleboard will stop movement and sound the alarm.