As indicated in the previous post, this week featured the first class session exclusively dedicated to project work. This provided an amazing opportunity for the Road Watch team to collaborate for an extended period of time. We met virtually, via Teams, to discuss project status and exchange ideas. It was a nice change of pace to be able to work on a segment of a project and be able to immediately ask a teammate for their input, instead of having to message them and wait for their reply. Several developments came out of this meeting. First, we began planning a design for a physical LiDAR testbench, which would use extruded aluminum bars to create a track for a scaled-down car. Next, we discussed the status of the Detection Subsystem PCB, which was just ordered, and will hopefully be arriving next week. Finally, several team members configured their local Git repositories, so that they could track their code changes and commit to the team repository on GitHub. Overall, the meeting was very productive. QRB2 is rapidly approaching, so be sure to read next week’s post to follow our preparation efforts.

Based on feedback from QRB1, the Road Watch team decided that it is imperative to have a fully working prototype of all non-AI aspects of the project. This essentially means everything must be working in an integrated system, with the exception of the predictive algorithm and the computer vision system. Thus, substantial progress has been made this week towards this goal. For instance, Darrion and Skyler successfully implemented the 1D to 2D LiDAR conversion process, and have all mechanical, electrical, and software components working in tandem to provide a 2D scan of an area. Will also made substantial progress on the Detection Subsystem PCB, and will be ordering it later this week. Billy and Rolando have been busy sourcing many of the electrical components required for our system. Finally, Evan was able to extract vehicle trajectories from the NGSIM dataset, which will be necessary for training the predictive algorithm. Next week, we have been given the whole class period to do more project work, so be sure to check in next Friday to see how our collaborative prototyping efforts pan out.

This week, the Road Watch team delivered its first Qualification Review (QRB1). As discussed in the previous blog post, the QRB events provide an opportunity for teams to receive constructive feedback and a risk assessment from a panel of experts. Our feedback was mostly positive (which was good to hear), though there was some issues raised. First, the panel though that some elements of our system (presumably our computer vision solution and our predictive algorithm) were ill defined at this point, which is a point that we also raised. Second, they indicated doubt in our ability to satisfy all design specifications with our prototype, probably because of the uncertainty of the viability of our predictive algorithm. Further, the panel recommended that we focus on creating a minimal viable hardware solution before we proceed to optimize the predictive algorithm, and they emphasized that this should happen soon. Thus, we are ramping up efforts in our design work for these hardware deliverables. Be sure to come back next week to see the progress we’ve made.

This week, the Road Watch prepared for its first Qualification Review Board, or QRB1. At a QRB, our team will present our project to experienced faculty, who will offer feedback on the project. The main objective is to pinpoint risks in the project, and to devise mitigatory efforts. In the context of product development, a risk is an either anticipated or unanticipated factor which might hinder the development of the product. An example of a currently identified risk is that binocular vision’s increasing uncertainty with distance could render the computer vision approach to this project unviable (see the figure below for a more in-depth explanation). In the QRB, our team plans on giving a quick overview of our project, and then each member will present their current progress and discuss associated risks. Check in next week to see what feedback we got for our QRB1.

This week, the Road Watch Team had its first in-class work session. These are a new feature of IPPD 2 that were not offered in IPPD 1, as IPPD 2 focuses primarily on the practical realization of the system (which necessitates collaborative in-person work). It was very nice to have all team members physically together in one place, all working on different parts of the system. This provided a good opportunity to quickly exchange ideas between teammates and to catch up on everyone’s progress. Director Latorre indicated that these in-class work sessions would be a bi-weekly occurrence, which is greatly appreciated by the Road Watch Team. Currently, work is distributed between teammates as follows. Darrion is developing the LiDAR driver for the Jetson Nano. Evan is working on identifying predictive machine learning algorithms. Will is evaluating the camera throughput on the Jetson Nano. Rolando is identifying an IMU to assist in aligning the system. Skyler is designing the system enclosure. Finally, Billy is identifying power supplies for the system. Be sure to follow future posts to see how each group member progresses in their tasks.

Welcome to the New Year! In our first week this semester, after a refreshing break, the Road Watch team began with a lecture by Professor Bill McElroy, P.E.. He focused on the topic of credibility, and showed how it was composed of credibility and competence. Then, we began planning out the upcoming semester, by creating a dependency graph and estimating the duration of each major milestone. Our critical path, or longest chronological path through the dependency graph, takes 72 days to complete. It should be noted that this is a hastily constructed draft, and the actual semester may take a massively different course. However, the 72 day critical path should comfortably fit into the spring semester with spare time to make some mistakes. Keep checking back weekly to see how well we stick to this plan.