Constraints - achyut3598/SmartHackSmasher GitHub Wiki

As with all projects, there are a number of constraints that limit our scope and deliverables. These are listed below.

Economic Costs - The economic cost for Smart Hack Smasher is rather low however there are some potential constraints financially. One is that we cannot afford a driverless car to collect data for this project as they are prohibitively expensive. So instead we will settle for freely available datasets. Another potential cost is hardware that will be used to demonstrate the project. Instead of using an actual car system, due to budget constraints, we will have to settle for a Raspberry Pi running a simulated car.

Time - The maximum time we get to complete this project is about 9 months cumulatively starting from the fall term till the end of spring term. Since this is a research project, there is a lot of trial and error involved, so having a certain time limit within which all the goals for the project need to be accomplished does make things a little tricky and depending on how our trials go, we may or may not have to reduce the list of achievable goals for the project.

Scope - The team agrees on the scope of the project. We all understand that since this is a research project, we have to move one step at a time and the success of the project depends upon how successful our initial modules were. Due to the time constraints, we have picked only 3 car modules to work on that are - Road Sign Detection, Object Detection, and the GPS module. Since we don’t have an actual self-driving car, this project is mostly theoretical.

Professional/Technical - This project features research mainly focused on the area of Machine Learning, which is a developing field. The lack of a concrete “best solution” in this area and our group being composed of college students will restrict how far we can go in this project with our level of expertise.

Ethical/Legal - Because self-driving cars do have real world impacts on the health of individuals using them, software regarding self-driving cars will need to be made with the safety of the drivers and passengers at the forefront. Therefore we cannot make any software that prioritizes security over safety.

Security - There are few security concerns within the scope of the project. As Smart Hack Smasher is a proof of concept research project, it will not actually be deployed on the roads. If however it was deployed on the road, a driver’s driving history would have to be securely kept. The type of data that is used for Smart Hack Smasher such as GPS and camera data can be used nefariously by any bad actor that would acquire it.

Social - This project will impact the lives of any who use or consider using self-driving cars. Because this is a data-focused project we are constrained on the amount of data we can record/log from a consumer privacy perspective. We are also constrained by the “public view” of self-driving cars and their safety.

Environment - For this project there will be no environmental concerns as Smart Hack Smasher is intended to be a mostly theoretical, demonstration project. In the case it was actually fully developed beyond the proof of concept phase, it could impact road conditions. This may lead to more efficient traffic which could reduce energy consumption on vehicles.

Diversity/Cultural - The dataset we will be using for traffic signs is the German Traffic Sign Benchmark, so our results will be constrained to information garnered from that culture. Additionally, different parts of the world have different perspectives on what is considered “safe” and “acceptable” which we may have to consider when deeming the approach a success or failure from a security standpoint.