Mercury Mesh: P2P Networking for Autnonomous Vehicles

[Sam3077]

Hey everyone!

Our team, Mercury Mesh, seeks to create a P2P networking infrastructure to enable autonomous vehicles to communicate and use information from nearby vehicles.

View our website at mercurymesh.dev

Objectives: Adapt algorithms similar to torrent and WAP protocols to enable a communication channel. Make use of static base-stations to administrate and bootstrap the network when necessary and to aid in calculating the local position of the vehicles. Create a generic data structure to allow different data to be shared in a vehicle and technology agnostic method. Adapt autonomous algorithms to make good use of the data.

Challenges: The network must be highly adaptable and fast since the vehicles are in transit and won't remain in-range for long. We must be able to create a networking protocol that is fast yet reliable and that reduces interference between multiple vehicles. We have to figure out how to incorporate shared data into a vehicle's current map of it surroundings.

Current Team Members and Their Duties: Benjamin Wadsworth: Positioning and Networking Sam Hirsch: Modeling and also Networking Dalton Clift: Robotics and Sensors and also Networking

Joining Mercury Mesh: We are looking for team members who are excited by the idea and are willing to put in a good amount of work. Helpful skills would include:

Embedded Systems / Robotics Networking (specifically with P2P networks) Computer Vision and Environment Detection

Contact: You can reply directly to this issue or contact Sam Hirsch at sam.hirsch66@gmail.com or on Discord at hirschy#4769

[TristanStotesbery]

Group 3 (Tristan Stotesbery, Spencer Durrant, Ben Van hoose): This project seems good because it improves driving safety and has potential to save lives. It seems like it will be very expensive to implement, and will have to be thoroughly tested, as safety is extremely important when it comes to autonomous vehicles. The network will need to be very secure to protect from attacks. The marketplace for this project seems very wide, including car manufacturers, businesses, and everyday commuters. Out group would consider buying this as a product if it was proven to be completely secure and safe. We would improve this project by clarifying what parts we are working to automate ((Traffic flow, steering, power management, navigation?)

[JstaNate]

Comment from Group HealthyBois (Nathan Hummel, Brady Smith, Michael Linnebach):

What is the practical use?
-  This would make traffic a lot better and safer in the long run.
What are the challenges we think of to implement this project?
-  Implementing a realistic 3D environment simulator for testing.  How will you make it as accurate as real world.  How much 
   development time will this take and how much time will it save?
-  Are you going to build your own RC car or retro-fit one already built.
-  Are the autonomous control algorithms free?
What is the marketplace for this project?
-  Any car market.  Truck fleet companies, ie: walmart.  
Will we buy it?
-  Probably not on a large scale as we dont own fleets.
What to improve from an outsider's perspective?
-  Be able to have a backup plan that is an acceptable ending point if the project is to large/complex to finish.

[salwab]

practical use is to facilitate the safety of driving vehicles. some of the challenges would be testing those simulations in the road while making sure the safety of drivers. marketplace: all drivers. yes *what to improve? clarify the instructions in their simulator.

[altonbill]

Jacob Bills & Collin Pollard

What is the practical use of this project? Improving vehicular safety and reducing traffic

What are the challenges to implementing this project? Reliability needs to be 100% given the use cases Unlikely to compete with commercially available alternatives

What is the marketplace for this project? Car Vendors Government Agencies

Will we buy it? Probably not, unlikely to achieve enough adoption to be usable on a wide scale

What to improve the project from an outsider’s perspective Ensure interoperability so it can interact with other platforms

[jason-zavala]

What is the practical use of this project? Saving lives by increasing vehicle safety

What are the challenges we think of to implement this project? reliability is the largest concern here. You're going to need 100% compelling, thorough, and convincing testing.

What is the marketplace for this project? Everyone!

Will we buy it? For sure! If you get a 100% autonomous vehicle working let me know.

[MrYo531]

What is the practical use? P2P networking for auto vehicles. Can communicate with other cars and improve traffic and safety. What are the challenges we think of to implement this project? Software algorithms seem pretty complicated Managing the network protocol can be difficult, and same with the electronics. potentially interested rocketeers. What is the marketplace for this projet? Anyone who owns Tesla or other autonomous vehicle Will we buy it? I don't have a Tesla What to improve from an outsider's perspective? I'm curious how much this costs, and whether its a straight forward process.

[alexCharters]

Comment from Group 6 Emmanuel Lotubai, Brandon Chen, Alex Charters:

What is the practical use? To make roads safer.

What are the challenges we think of to implement this project?

• Testing on fully autonomous RC cars.
• Steady and reliable communication.
• Affordability.

What is the marketplace for this project?

• People who already have autonomous vehicles.

Will we buy it? Nope. We're broke.

What to improve from an outsider's perspective? Typos in website.

[Huchimama]

vecterU (Scott Crowley, Aaron Tea) This has many practical uses and could be the holy grail of autonomous driving if successful. I'm sure you are well aware of the many challenges involved especial those that at arise from driving at high speeds and weather conditions. There is a huge market for this and we would buy if it worked. Our biggest concern would be security. Bad things could happen if a malicious user was able to take control. We are looking forward to seeing the finished product.

[Interestle]

Comment from Group 9 (Colton Watson, Seth Jackson, Phelan Hobbs):

What is the practical use? Too many accidents occur every year, with far too many fatalities. This would help reduce that.

What are the challenges we think of to implement this project? Wireless communication is challenging, especially when the communication channels grow in density. It looks like there will be a large amount of data being transferred, maintaining that would be challenging. Interpreting the data could also be troubling. FCC stuff, and interfering with other wireless communications.

What is the marketplace for this project? City Planning, car manufacturers, anyone with a drivers license, or drivers without one.

Will we buy it? Invest, yeah. If it was already in my car, even better.

What to improve from an outsider's perspective? Maybe try mitigating the risks? This is a really ambitious project.

[lrm77]

Comment from Group 5 (Luke, John, McKay, Benjamin): --Practical Use Allowing communication between autonomous vehicles to improve driving safety. --Challenges -The main challenge will be maintaining reliable communication between vehicles moving at high speeds. Especially with the high volume of data that needs to be transferred to make the product useful (e.g. information about surroundings and environment), it will be challenging to maintain a high enough data transfer rate. -Setting up a realistic test environment. --Marketplace This project would be most useful for manufacturers of vehicles who can implement the communication system. --Purchase? No, this product is not currently applicable to us, but in the future it could be a useful product. --Improvements? -Possibly test the communication system before implementing the RC cars autonomous functionality.

[dclift97]

Thank you for all your comments and feedback regarding our project. Security and reliability are concerns that we are already looking into. To address some common misconceptions on our project we have a few things to note:

First - This is primarily a proof of concept using RC cars and simulation. Our customer is primarily manufacturers and other autonomous vehicle researchers. You should invest because they have deep pockets

Second - Security will definitely be a large focus throughout this project, there is some built in encryption with the V2X chips, that we are thinking of integrating alongside our own security protocols.

Third - Our focus is the communication and algorithm adaptation. Steering, power management, navigation, etcetera, are handled by the pre-existing autonomous systems that we will integrate with. We may focus specifically on traffic flow to add newer co-operative traffic planning. The goal is to share data and allow vehicles to make decisions based on sensor readings from surrounding vehicles.

Fourth - We are aiming to extend existing V2X implementations, specifically by having cars share sensor data in a uniform manner between different manufactures and sensor types. ​This is not a closed environment, rather it is an open addition to available products.

Fifth - 3D environment is provided by a Microsoft project (AirSim) through the Unreal Engine it is primarily a testing/prototyping environment.

[YeZhou0226]

• Is the prototype related to the project? (with comments) Yes. Auto-range with LPS is important for keeping track of cars' location. • Is the prototype convincing for you to invest? (with comments) Yes. The prototype contains basic concepts for their project.

[MrYo531]

Is the prototype related to the project?

• Having an RC car is a good idea. Cheap and gets the point across of being a vehicle.
• Communicating w/ base stations and other vehicles/RC cars, but this will still show proof of concept as well.
• shows autonomous driving
• the simulated environment is good because it can show off the whole software, with no hardware/real-world limitations

Is the prototype convincing for you to invest?

• Solid prototype, I would invest. It seems like this will scale well into the final product and catch any problem that would arise.

[Interestle]

Comment from Group 9 (Colton Watson, Seth Jackson, Phelan Hobbs):

Is the prototype related to the project? Yes, it’s the project, but scaled down.

Is the prototype convincing for you to invest? Probably not, but if the autonomous driving was good, maybe.

[Durrantula]

Group 3's response Is the prototype related? It's definitely related, but the autonomous functionality needs to be clarified.

Is the prototype convincing to invest in? Not yet, what specifically makes it a prototype?

[BradySmith1019]

HealthyBois comment:

Is the prototype related to the project? Yes, the prototype seems well-thought-out and detailed. The feasibility of the project is in question a bit.

Is the prototype convincing to invest in? If completed, definitely warrants investment.

[mfateh93]

Group 8 (Salwa, Josh, Todd, Mohammed)

Is it related to the project? Yes it is related to the project.

Is the prototype convincing to invest? It seems like a lot to do for a prototype, we are not sure if this can be accomplished with what is left during the semester.

[alexCharters]

Is the prototype related to the project?

• Yes. however, we think that the simulated environment may not be totally necessary to demonstrate your project's goals.

Is the prototype convincing for us to invest?

• Yes, if the RC car can reliably communicate with the base station.

[altonbill]

Is the prototype related to the project?

Yes getting a set of vehicles together that can sense their environment is important.

Is the prototype convincing for you to invest? No, I don't think getting the RC cars and simulations is going to be the hard part, getting some sort of V2V communication would be.

[bbleaptrot]

Comments from Team 5 (Luke, Benjamin, John, Mckay): -Is this prototype related to the project?

Somewhat. The goal of this product is to transmit the data between vehicles. From the prototype description, it is gathering the data, but it is not sending the data to another vehicle to process.

-Is the prototype convincing us to buy?

No. The prototype would need to also transmit the data and confirm it has been received good data from another vehicle

[ColinPollard]

Is the prototype related to the project? Absolutely.

Is the prototype convincing for us to invest? If you can pull off the stated goals, this is a great risk reduction step. Whether that is enough to convince investors for funding is a more difficult sell.

[Huchimama]

vectoru (Aaron Tea, Scott Crowley) The prototype is related to the project and easily understood. It is very descriptive and convincing. Nicely done!

[Sam3077]

Do people correctly understand what you plan to do? For the most part, yes! However, there seems to be some confusion on the intended goals of our prototype and how we will be proving our project. The primary goal of this project is to analyze and show the efficacy of inter-vehicle communication. To that end, we are not re-inventing the wheel by designing our own autonomous algorithms nor are we using hyper custom communication protocols. Our physical prototype, using RC cars, will use off-the-shelf LPS chips and V2V communication chips. As such, the majority of the difficult parts involved in communication will already be done for us! Additionally, the simulation environment aims to further examine real-world driving scenarios that we couldn't test in real life.

What do you plan to improve based on these comments? We will work to clarify more what specific technologies we plan to use and specify the scope of our project. We definitely hear the concerns about the large amount of work in this project, so we'll work to limit the scope of our project to make it easier to implement.