{"id":196,"date":"2016-10-21T21:20:18","date_gmt":"2016-10-22T01:20:18","guid":{"rendered":"http:\/\/mrsdprojects.ri.cmu.edu\/2016teami\/?page_id=196"},"modified":"2017-04-07T20:47:14","modified_gmt":"2017-04-08T00:47:14","slug":"implementation","status":"publish","type":"page","link":"https:\/\/mrsdprojects.ri.cmu.edu\/2016teami\/implementation\/","title":{"rendered":"Subsystem Description and Implementation<\/font>"},"content":{"rendered":"

\u00a0Spring Semester Updates<\/h1>\n

The two major components of the project for the spring semester are the Navigation sub system and the Estimation sub system. This section covers the progress made in both the sub systems.<\/p>\n

1. Navigation<\/h2>\n

1.1 Advanced Base Construction<\/h3>\n

We choose to go with in-wheel hover board motors for the motorized upgrade of the SoyBot. The detailed trade study can be found here<\/a>. Additionally we designed a new mount\/adapter for these motors. Next the entire assembly was integrated with the SoyBot chassis.<\/p>\n

\"\"

Figure 1.1 In-wheel motors<\/p><\/div>\n

\"\"

Figure 1.2 Motor Mount\/Adapter for in-wheel motors<\/p><\/div>\n

 <\/p>\n

\"In-wheel

Figure 1.3 The new motorized wheels have been mounted onto SoyBot with the new adapter<\/span><\/p><\/div>\n

1.2 Power Train<\/h3>\n

The new drive power train is also mounted on SoyBot. Each wheel is powered by one 36V 4.4 Ah battery. Each wheel also has its own motor controller board and arduino interpreter to decode messages and interface with the main vehicle computer.<\/span><\/p>\n

\"\"

Figure 1.4 Power Train for SoyBot<\/p><\/div>\n

1.3 Sensors and Calibration<\/h3>\n

The two sensors that will be used for navigation have also been fully integrated with the vehicle. We are using two lidars facing at a 45 degree angle down at the row of plants to assist with high level navigation planning. The three scenarios that we will be dealing with are shown on the right. The vehicle can be partially in the row where only the front lidar is used, fully in the row where both lidars are used, and leaving the row where only the back lidar is used.<\/span><\/p>\n

\"\"

Figure 1.5 Navigation Sensors for SoyBot<\/p><\/div>\n

1.4 Localization<\/h3>\n

This video shows a test run of autonomous navigation with one lidar through a 6+ meter long row of plants. We are using RANSAC and additionally filtering based on the geometry of the lidar position to fit 3 lines to the rows, shown in blue. We then use these three lines to get an estimate about where the vehicle is with respect to the center row, shown as the yellow arrow.<\/span><\/p>\n