Category: Uncategorized

Final-Report-ECE Capstone

Aditya Ranade

• What did you personally accomplish this week on the project? Give files or photos that demonstrate your progress. Prove to the reader that you put sufficient effort into the project over the course of the week.

This week, I worked on improving the computation time of the robot. We keep the map constant between determining frontiers and finding the trajectory. As a result, we get both the results with only one service call. This reduced our computation time from 17s to around 10-11s.

• Is your progress on schedule or behind? If you are behind, what actions will be taken to catch up to the project schedule?

We are finished with our project and just making minor optimizations to improve obstacle avoidance and reduce time taken.

• What deliverables do you hope to complete in the next week?

Final Report!

Akshat Jain

• What did you personally accomplish this week on the project? Give files or photos that demonstrate your progress. Prove to the reader that you put sufficient effort into the project over the course of the week.

This week we worked mainly on setting up a cardboard maze that resembles our demo space. We tested our robot multiple times in this maze ensuring the thermal camera works correctly and is able to detect people most of the time. We recorded all of these results to include in our final report.

• Is your progress on schedule or behind? If you are behind, what actions will be taken to catch up to the project schedule?

We are more or less done with the project and all that remains is further testing and recording of results.

• What deliverables do you hope to complete in the next week?

We hope to work on the final report in the next week.

Wenxin Xiao 

• What did you personally accomplish this week on the project? Give files or photos that demonstrate your progress. Prove to the reader that you put sufficient effort into the project over the course of the week.

This week we spent a lot of time figuring out how to set up the maze we are showing for public demo. We assembled a test space that resembles the demo space we planned and tested our robot many times to ensure the navigation and thermal are working as expected.

• Is your progress on schedule or behind? If you are behind, what actions will be taken to catch up to the project schedule?

We essentially finished our project but we intend to test and record our results with more environments and iterations to provide a fuller discussion for final report.

• What deliverables do you hope to complete in the next week?

We hope to present in the public demo and finish the final report next week.

TEAM UPDATE

This week, we were able to significantly reduce computation time from 17s to 11s. We also mask some of the computation time by spawning a thread to rotate the robot at the same time. We also set up our final demo environment using cardboard boxes to test our robot before the public demo. Overall, the success rate of completion and human detection is something we are quite happy with.

Aditya Ranade

• What did you personally accomplish this week on the project? Give files or photos that demonstrate your progress. Prove to the reader that you put sufficient effort into the project over the course of the week.

This week, Akshat and I found the bug in frontier exploration which did not appear in simulation. The issue was the robot would sometimes execute an older trajectory instead of the current one. We found that this was due to a concurrency issue in which the path was beginning to execute right before the trajectory was received. Now, exploration works very robustly.

• Is your progress on schedule or behind? If you are behind, what actions will be taken to catch up to the project schedule?

We are on schedule currently. The thermal camera has been integrated into exploration and detects people in periodic intervals. Frontier exploration is also giving us good results with about 8/10 runs succeeding in testing.

• What deliverables do you hope to complete in the next week?

I plan on spending time next week to speed up our computation times so we can reduce the total time taken by the robot to cover the entire map.

Akshat Jain

• What did you personally accomplish this week on the project? Give files or photos that demonstrate your progress. Prove to the reader that you put sufficient effort into the project over the course of the week.

This week, I helped Aditya fix the concurrency bug that we were facing last week which was due to the robot trying to execute a previous path before it received the complete trajectory. I also worked on the final presentation as well as getting data from testing to include in our slides.

• Is your progress on schedule or behind? If you are behind, what actions will be taken to catch up to the project schedule?

Frontier exploration has started working much better and executes correctly without collisions in our test areas most of the time. The robot gets quite slow when it is requesting a new path and we are hoping to reduce the computation time.

• What deliverables do you hope to complete in the next week?

We hope to reduce computation times as well as thoroughly test our thermal camera in multiple environments by the end of next week.

Wenxin Xiao 

• What did you personally accomplish this week on the project? Give files or photos that demonstrate your progress. Prove to the reader that you put sufficient effort into the project over the course of the week.

This week I successfully integrated thermal camera to robot exploration with help from Aditya and Akshat. I also worked on the final presentation and tested the robot in apartment hallways and dining room with teammates. Due to the limited range in which thermal camera can successfully identify humans, we changed our algorithm so that the lidar rotates every few seconds to scan the surrounding area for human. The updated logic is that the thermal camera consistently checks for image temperature and once a block of image beyond preset threshold temperature was observed it signals the lidar. The lidar contains information of the obstacle observed and we can calculate the distance and angle of the person observed from the position of the robot and mark it on a map.

• Is your progress on schedule or behind? If you are behind, what actions will be taken to catch up to the project schedule?

We are currently on schedule. The concurrency issue causing the robot to execute old trajectories repeatedly was fixed and we successfully integrated thermal camera into robot exploration.

• What deliverables do you hope to complete in the next week?

We hope to test our thermal camera thoroughly in multiple environments by the end of next week. If time permits, we hope to speed up computation times and retest on multiple environments.

TEAM UPDATE

This week, we fully integrated human detection and frontier exploration subsystems. The robot periodically rotates during exploration to scan for humans in its vicinity. We use LIDAR data along with the thermal camera reading to interpolate where a human is located. The equations can be found in our presentation.

We also worked on testing both systems in the corridor of our apartment building and a portion of our apartment. We recorded preliminary test results for the final presentation.

Aditya Ranade

This week I worked on writing code for frontier exploration. I have working code for detecting frontier cells – which are cells at the boundary of unknown and empty regions of the map. It is often the case that frontiers lie in clusters, as a result, we find the center of the cluster and make the robot navigate to it. Now that we have found frontiers, we are working on a path planning and obstacle avoidance algorithm so that the robot can explore autonomously.

Although we got a significant portion of frontier exploration working, I was hoping to also finish path planning. However, I was extremely busy with Greek Sing and other assignments this week.

Over this weekend and the beginning of next week, I will wrap up path planning and have a working frontier exploration subsystem.

Team Report

This week we finally received our iRobot Create 2. We were able to send it velocity commands and get it to move and rotate. We will be 3D printing a frame to mount on the robot to accommodate our LIDAR, thermal camera, and RPi. We also made a lot of progress on frontier exploration.

For the demo next week, we plan to show a working version of our frontier exploration subsystem in simulation. We will also show that we can extract range data from the lidar as well as thermal camera images. After the demo, our goal is to integrate all these components and move from simulation to hardware testing.

 

Aditya Ranade

Over this week, I worked on incorporating feedback from the presentation into the software subsystem design portion of the project for the Design Report. I am in the process of writing a detailed summary of all the nodes involved on the software side of the project. In particular, I had to expand on our usage of scan matching and particle filter as the level of detail for the presentation was not sufficient.

I also worked on finalizing the our block diagram including all the message types that are communicated between nodes as well as a legend (which was pointed out by the Prof. Tamal after the presentation).

I also began working on setting up the Scan Matching and Particle Filter Nodes to verify the accuracy of maps generated by manually controlling the robot.

We are slightly behind on ordering parts for the project. We will be ordering the robot chassis, IMU and LIDAR this Monday and assemble them over spring break.