This project is the Senior Design group 11 for the University of Central Florida College of Electrical and Computer Engineering
Team: Josh Franco John Millner Jeff Strange Richie Wales
SigSent is an intelligent, multi-terrain hexapod robot built for robust, reliable security solutions. Using an evolved Neural Network using NeuroEvolution of Augementing Topologies (NEAT), SigSent determines when the road is getting rough and automatically transition from a four-legged drive mode into a walking mode using all six of its legs. This allows base operators to work at a higher level, simply providing GPS routes to the robot and leaving it to figure out the rest itself.
SigSent's gait is generated using a Genetic Algorithm to find the most optimal, stable walking motion under its given load.
SigSent features a computer vision module that runs a basic pedestrian detection algorithm so that base operators can be notified if an anomalous figure has appeared that needs addressing.
The base operator can also TeleOp with a provided joystick, listen to audio from the robot's integrated microphone, and also use a headset of their own to output audio from SigSent's own speakers for two-way communication. Diagnostic information regarding things like battery life, servo pose information, and CV/NEAT outputs are provided as well in the GUI application. The interface was created using PyQt5, Python bindings for Qt5.
This robot runs on a raspberry pi 3 uses the following sensors: Sparkfun Razor IMU 9DoF 14001 Sparkfun Venus638FLPx RasPi Cam Hokuyo UTM-30lx LIDAR
To begin the robot first launch roscore on a basestation then launch local.launch from the Raspberry Pi followed by launching remote.launch from the basestation
ensure that environment variables such as ROS_MASTER_URI and ROS_PI are properly set and that the package is properly sourced on both computers source ~/SigSent/catkin_ws/devel/setup.bash ROS_MASTER_URI=http://192.168.1.110:11311 ROS_IP=192.168.1.110
Sensors can be easily visualized using RVIZ rosrun rviz rviz once launcheed rviz should use the rviz configuration file found in the root directory of this github