Posted on

Team Status Report for 03/26

On the software side, we encountered issues that http requests cannot be received by Pi. After testing, we found out that port forwarding is preventing the http requests from reaching Pi. We’ll be looking for solutions to circumvent this issue. (Detail in Roger’s report). On the computer vision side, we’re able to setup camera on the raspberry pi to run OpenCV algorithm. On the sensor side, the communication between ultrasound sensor script and the software is finished, and the runtime of parking spot availability is reduced. This allows us to quickly detect entering/exiting of cars from the spot, and later providing navigation to leave the garage once a car leaves.

We’re currently behind schedule due to the modification of design on the server side. We’ll try to resolve this issue by next week.

Posted on

Yuxin’s Status Report for 03/26

This week I’ve been working on the I/O between the ultrasound hardware component and the software component. The ultrasound python script is able to communicated with the software system about the state of parking spots (1: Available, 0: Taken, -1: No sensor). The runtime of distance estimation is also reduced from 2 sec/sensor to 0.1 sec/sensor by multiprocessing and settle-time reduction. Meanwhile I’m working on the design of LED & Raspberry Pi circuit and scripts. I’m on schedule for this week, and I hope to finish the LED controlling part next week.

Posted on

ruizhezh’s Status Report for 03/26

What did you personally accomplish this week on the project?

Apart from the ethics assignment, I was trying to receive HTTP requests from AWS servers on Raspberry Pi. 

However I discovered that there are many technical difficulties with this process, that if use __ to connect Pi to the internet:

  1. mobile hotspot: HTTP requests cannot be sent to Pi as port forwarding are not supported / cannot be configured. 
  2. CMU Wifi: For whatever reason we are still uncertain about, the Pi is unable to connect to CMU wifi and all online answers have failed to rectify this problem. Even if it can connect to CMU wifi, port forwarding will still prevent the HTTP request from reaching PI. 
  3. A WIFI router: we cannot find a ethernet port to set up a router in the parking garage. 

Researching this topic and testing the above solutions (without success) mostly summarizes my work this week. 

Is your progress on schedule or behind? 

My progress is behind due to the above issue. I am making changes to the design to rectify this issue. Hopefully the issue can be solved soon. 

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

Planning to test ways to circumvent the issues posed by the Raspberry Pi IP addressing issues; 

Planning to link together Raspberry Pi and AWS Server for communication testing. 

Posted on

Ke Xu’s status report for 3/26

What did you personally accomplish this week on the project? 

I finished the camera and pi setup and worked on the similarity score algorithm.

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

My part is on schedule. But we are having issue with the http request on pi. Our pi can send http request but cannot receive the content from the request. Roger may need some help in the http request.

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

Finish the similarity score algorithm. Help Roger with the http request and prepare for our interim demo.

 

Posted on

Ke Xu’s Status Report for 03/19

I have finished plate recognition algorithm with OCR. I have tested the algorithm and have it installed on our pi. Next week, I will have the camera connected to pi and test the algorithm in garage to prepare for our demo.

Our plate recognition algorithm may not recognize every letter on the plate correctly, but we can calculate a similarity score to differentiate different cars. I will implement this next week.

Posted on

ruizhezh’s Status Report for 03/19

What did you personally accomplish this week on the project? 

  • Completed prototype of server-side management application; 
    • The server is now capable of receiving Json formatted RESTful HTTP requests including: car-entering-entrance, car-entering-parking-spot, car-leaving-spot, car-leaving-parking-lot; 
    • Items required for navigation is completed; 
    • [TODO] Parking lot data import is currently being worked on; 
    • [TODO] Most of the code surrounding communication with RBPI is done. The exact formatting of the request will be tested next week; 
    • [TODO] OpenCV plate ambiguous matching is isolated and progress-blocked by XuKe’s OpenCV implementation; 
  • Worked with XuKe to setup Camera and Python environment for his RBPI; Worked on brainstorming solution to inaccuracies in OpenCV; 

Please refer to the Github repository link for source code updates; 

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

Progress is a little behind due to needing to help XuKe with his OpenCV implementation: 

  • Caused changes to the management software in anticipation for ambiguous plate information; 

We are hoping to make progress on OpenCV as soon as possible to solidify design specs for the software. At this state, we are still not certain on the specs of our software interface due to not knowing: 

  1. How long OpenCV will take to process;
  2. How accurate OpenCV will be able to recognize the plate information; 

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

  • More testing on the RESTful API with AWS servers; 
  • Parking lot data import is currently being worked on; 
  • Find out how to perform HTTP communication between Raspberry PI and AWS server; 
  • Connecting the sub-systems for demo; 
Posted on

Team Status Report 03/19

On the computer vision side, we’re able to run our OpenCV model on raspberry pi for plate recognition (detail in Ke’s report). The plate recognition is not 100% accurate (there can be addition/deletion/mismatch of letters in the result). To address this issue, we plan to match cars to plate numbers by comparing similarity score.

On the server side, the management part is mostly finished with blanks left to fill during testing and OpenCV integration. The design of navigation software is finished, and the design of data importation is ongoing (detail in Ruizhe’s report) . Race conditions may be possible for HTTP requests coming in at the same time. Solution to this is still under investigation but should be manageable.

On the hardware side, multiple sensors were successfully connected with Raspberry Pi to measure distance with adequate accuracy.

There’s no major modification of the system design, and we’re currently on schedule.

Posted on

Yuxin’s Status Report for 03/19

Personal Accomplishment
This week I’ve been working on the parking component — the ultrasound sensors for detecting which spots are taken. I designed the circuit to allow 3 ultrasound sensors be wired with the raspberry pi. My python script running on Pi can detect the distance measured by each ultrasound sensors with a resolution of ~4 cm. The threshold for detecting a car is when distance measured is <=40 cm, and this can be tuned later during testing. So far the 3 sensors is a proof of concept, and I can add more sensors in a similar way.

Personal Schedule
I’m on schedule.

Next deliverable
For next week, I hope to complete the navigation LED component at intersections. This will involve LED strips, Arduino, and Pi from the hardware side. Instructions from the server will need to be received by Pi and Pi will control Arduino to light up corresponding LED strips to show directions.