Adnan Khan and Noah Borel
Over the course of the next year, we will be working to create an anti-drone defense system. We, Team 2, will be working on the vision segment of the system while Team 3 will be working on the physical intercepting segment of the system. As we work on these parts of the project, the final product will be able to efficiently spot a flying drone and incapacitate it safely without damaging it or causing a hazard to any nearby people (or buildings).
In the real world, as drone become increasingly used and put in the public eye, the market for an anti-drone system is very high. With many cases against drone fliers over privacy and trespassing concerns, an anti-drone system that does not damage the drone will be ideal. Usable by the government, private companies, and even for private housing, this product has great potential.
- The original goal does not require "not damaging the drone" (it's good to have, but not required), but not damaging the buildings or hurting people in the intercepting process. This goal has implications to the material and speed of the projectile.
Vision System: Team 2: Adnan Khan, Noah Borel.
Intercepting System: Team 3: Eduardo Guzman, Henry Hoare.
Groups will stay in touch via email and through mobile phone. At first, the two groups will develop each system individually. Once the systems become developed, the two groups will need to converge to integrate the two projects.
- Between team 3 and your team, you should define the hardware/software interface between your two systems clearly early on. Otherwise, it won't be possible to integrate two systems together at the end! It includes things such as
- architecture: are both systems controlled by a single microcontroller, or both of them have their own microcontrollers and communicate to each other through some communication channel (and what's the physics communication channel?).
- What are the data (e.g., distance, 3D angles), clock, and control signals between the two system?
- What are the timing requirements (e.g., target location data needs to be updated every 1ms) for the communication?
To create the vision system, we will need full understanding of several topics:
- Physics: Optics and the mechanism of the eye. (especially the optics of compound eyes and how to construct them.)
- Biology: Types of eyes found in nature and how they work.
- OpenCV: Understanding how to operate the program.
- You will need basic knowledge of image processing, pattern recognition, computer vision, motion tracking in the compound eyes context. The OpenCV probably cannot cover the compound eyes function. You need to develop algorithms to handle it.
Currently, we have basic knowledge of these topics. With this basic understanding, we have various ideas of what type of "eyes" we should use for the vision of the defense system. Yet, we don't know what we are capable of until we have learned all that we can. Our learning process has just started and we plan to absorb a lot before we begin diving right in.
As we expand our knowledge, we will know more materials that we will need. As of now, we need a Linux computer to run the OpenCV and later we will need a drone prop for our vision tests.
- You need to focus on the compound eyes mechanism as soon as possible since it may imply the needs of buying camera(s) and optical lenses, and the platform requirement to run the algorithms.
- If you consider to involve infrared vision, we also need to buy the infrared camera.
- Your research should not limited to "the materials and resources that are given" since we will purchase equipment/materials as needed (if affordable).
No comments:
Post a Comment