This week, our team went on-site to the FDOT Operations Center to collect data from our radar sensor. We plan to use this data to improve our hardware setup to maximize the range and accuracy of vehicle detection, as well as improving our software to maximize accuracy and make strong predictions of each vehicle’s trajectory.

The videos below demonstrate a vehicle intrusion during a lane closure, where the test vehicle first heads in the opposite direction on the correct side of the road, and then comes into the lane closure on its way towards the radar. Each video can be played side-by-side to watch the data collected on the radar as the test vehicle travels in real-time.

This week, we were able to start gathering data from the radar device and visualize the data on a graph. Below is a video that demonstrates how the radar picks up an object from a distance. Note that the orange dots represent a point on a object that the radar detects.

In the video above, you saw a spike with a lot of orange points and that represents a cardboard box that we move in front of the radar. We used this demonstration to provide clarity in front of eight judges whom provided feedback during prototype inspection day.

In all, this week proved to be a big step in our journey to create a device that alerts workers of traffic hazards ahead of time to save lives.

This week we finally got to interfacing with our radar module. We had a couple of set backs that have delayed us fully interfacing with the radar module so although we can now flash it we are unable to read data from it. To remedy this we ended up purchasing a second UART to USB adapter and once it arrives we should be able to read the radar data.

We also went out to FDOT operational facility to start some initial testing. Because we could not fully interface with the radar it was not as fruit full as we would have liked it to be but we did get some additional documentation about lane closures from FDOT while we were there. Overall, even with these setbacks we should be able to achieve our goal by prototype day.

Coming out of the PDR presentation last week, we have a clearer and more thorough understanding of when and in what instances our alert system would be used. Taking these and other points into consideration, we have begun solidifying our prototype’s technical details. For one, we have received our mmWave radar sensor to kickstart testing. We have also met with our sister team Awareables and determined our shared communication medium.

With that, our hardware and software subteams have started development for the upcoming prototype inspections. We look forward to onsite testing at FDOT’s facilities next week.

This week, our team presented our preliminary design review at the Florida Department of Transportation’s Gainesville Operational Facility. We presented the initial stages of our product to our liaisons and many other FDOT engineers. The experience was incredibly valuable, and our audience gave us great feedback and design considerations. Moving forward, we will take care to ensure that our product handles the needs of both lane closure and shoulder closure work zones, and we will prioritize ease of use in our design so that setup is frictionless for workers.

After our presentation, we also took a tour of the facility’s testing field. As we begin development in the next few weeks, we will coordinate further with FDOT to gain access to these testing grounds and additional equipment like tripods and stakes.

This week Road Response has refined the ideas we included in our Preliminary Design Report. While refining these concepts, we were able to come together on many questions that were crucial to our project’s success. How far away would a vehicle need to be for our solution to alert workers in time? How could weather conditions and traffic density affect our sensors? How will data from the device be captured and stored? These were all questions that our team discussed in our many meetings this week. For an example, if we want to detect a incoming vehicle going 65 miles per hour, we would need to alert workers of the approaching vehicle within 300 meters to give workers 10 seconds to evacuate. The figure below shows the relation between distance away and time given for workers to evacuate for vehicles going at similar speeds.

As our team weighed the pros and cons of each of the concepts we generated, our team is investing our time towards building a solution using radar to detect incoming vehicles. Radar is able to preform better in weather conditions and detect vehicles from a longer distance compared to some of our other concepts. Moving forward with this concept, our team defined our preliminary architecture for the hardware and software of our solution. Our hardware includes two radar sensors that provide data to a main controller. With the input received, we hope to use existing algorithms to find the speed, direction, and jerk of vehicles to determine if they will pose a threat to workers. Our team plans to continue developing this concept for our final PDR report and presentation.

This week, our team dedicated our efforts towards drafting our Preliminary Design Report, putting all of our research and work to date into a single document. On Wednesday, we got together to discuss our ideas and came forward with multiple plausible concepts. We’re excited to narrow down these concepts and possibly even combine them. For example, our team is looking into using two separate ideas, radar and LiDAR (light detection and ranging), to map the construction area in 2D and then detect objects intruding that area in 3D. We look forward to presenting the PDR to our sponsors, liaison engineers, and related officials at FDOT soon!