Forums

Will autopilot two hardware support V2V and V2I?

Will autopilot two hardware support V2V and V2I?

Will Tesla autopilot two hardware support V2V and V2I?
I understand that there will be hardware to support for WiFi but the question is will it support the 802.11p protocol. This protocol will allow Dedicated Short Range Communications ( DCRC )There has been much research on this technology over the past several years and the government has played a very big role in this research. I am not an autonomous vehicle engineer but I cannot see how full autonomy (level 5) can be achieved without this type of communication.
I have not seen any mention of v2v and v2i in any of Tesla’s documentation of autopilot 2 hardware capabilities. I'm sure Elon and his incredibly sharp engineering staff have contemplated this but it sure would be nice if someone from Tesla would come out and state whether or not they are going to have v2v and v2i support.

ReD eXiLe ms us | 07. August 2017

My guess is that their own Tesla Network for Transportation as a Service may well incorporate some form of Vehicle-to-Vehicle communication. Perhaps not directly, but possibly via the network allowing each car to know information based upon where what other Tesla products nearby have gathered. So, their own internal, proprietary Vehicle-to-Infrastructure system. Once that is proven to work, then Tesla may offer it as a solution for others. But on most projects, I expect Tesla to operate as would 'The Little Red Hen'.

Tâm | 08. August 2017

@pbender3469

Those protocols are great:

If my car can communicate with traffic lights, it can adjust its speed to run a long stretch with just-in-time green lights.

However, I think the difficulty are:

1) the agreed protocols among all parties.

2) The next is funding.

How to get funding to retrofit the traffic lights...

How to get funding to provide wi-fi hotspots....

3) Privacy: Some do not want to be tracked even anonymously.

bp | 08. August 2017

It seems to make sense that regulatory approval for FSD may require all FSD vehicles support and participate in V2V and V2I.

Has Tesla made any comments on V2V and V2I?

Tâm | 08. August 2017

@bp

No. Tesla is working on its Self Driving suite which has no V2V and V2I.

So far, 3 states have approved driverless cars. They only ask that those cars without a licensed human drivers must obey all traffic laws.

There's no requirement for V2V and V2I.

Florida
https://qz.com/781113/how-florida-became-the-most-important-state-in-the...

Michigan
https://www.cnbc.com/2016/12/10/michigan-lets-self-driving-cars-on-roads...

Georgia
http://www.ajc.com/news/local/georgia-officially-oks-self-driving-cars-p...

carlk | 08. August 2017

Is there V2V and V2I standards already? I kind of think Tesla would have incorporated them in current models if there are. They will just make FSD whole lot better whether they are required or not.

Tâm | 08. August 2017

@carlk

I think those standards are still in its infancy just like there are many promising battery breakthroughs that will make Lithium Ion battery obsolete.

The problem is cost!

It might cost about $600 billion–$1 trillion for next 20 years.

https://cei.org/blog/obama-highway-administration-released-troubling-veh...

That's why Tesla is still doing Lithium and conventional Self-driving suite until those other technology will come down on costs.

Dac | 08. August 2017

I do not believe V2V / V2I are required for full autonomy. Its not like you as a driver can talk to the driver behind you or in front of you. (Flipping him off doesn't count :P). This isn't to say that it wouldn't be beneficial, because I think it would.

carlk | 08. August 2017

I think V2V or V2I are of little use for now or for a long time to come. You can't have FSD to rely on it unless if every car and every road is equipped that way. Not sure if that will ever happen.

chris.pribe | 08. August 2017

While V2V R&D may not have been quite the YUUUUGE white elephant that Fuel Cells cars were, it is unnecessary for FSD and has only one likely near term use: Combustion car OEM's may promote mandated V2V to unknowing bureaucrats so as to create a tar pit to stall Tesla's progress. The most likely result of this would be to hand the self-driving technology business to China.

I'm hoping our policymakers are not quite that dim, nor still quite so in thrall to today's big energy and transportation players. There may be niche uses for V2V (e.g. supply line truck platooning), but far better to give over DSRC spectrum to some other use than to allow mandating V2V to hobble FSD progress in the marketplace.

ReD eXiLe ms us | 08. August 2017

Most of what I have seen and heard from the NHTSA notes that they want to take a Proactive, rather than Reactive, approach to handling such technologies. From fully autonomous vehicles to those that communicate with each other as a fleet. The NHTSA wants to ask the relevant questions ahead of time, before such technologies are widespread, so that they can have useful regulations and template legislation in place to deal with it all ahead of time. I think that approach is in place to foster such technologies for the sake of safety, as opposed to protectionism.

Feds: Cars must be able to talk to each other | USA TODAY
Nathan Bomey, USA TODAY Published 10:48 a.m. ET Dec. 13, 2016
www.usatoday com/story/money/cars/2016/12/13/self-driving-cars-vehicle-to-vehicle-technology-v-to-v-nhtsa-autonomous-cars-v2v/95367310/

--------

The text below is from a 4-page PDF file I found on the NHTSA's website...

THE ROAD AHEAD, National Highway Traffic Safety Administration Strategic Plan, 2016—2020

Message From the Administrator

At the National Highway Traffic Safety Administration (NHTSA), where our mission is to save lives on American roads, we often talk about two very important numbers: 35,092 and 94.

In 2015, 35,092 people lost their lives on our roads, a 7.2-percent increase over 2014. Those aren't just numbers. These were our children, parents, spouses, friends, neighbors, and colleagues. They were somebody’s everything.

The second number is 94. That's the percentage of crashes caused by human choice or error. That's a fatal decision to get behind the wheel after drinking. It's the decision to speed through an intersection as a light is changing, or the inability to brake fast enough to avoid the person who just sped through. It's the decision to drive after a sleepless night or to send one more text from behind the wheel to let someone know you're on your way.

These two numbers are a call to action, because, frankly, at NHTSA we are working to create a future where no one dies on our roads. How then, do we get to zero fatalities from a near-term record high? We know we can't do more of the same and expect a different result. This plan is our renewed commitment to our many successful safety programs, and our dedication to finding innovative tools that save lives.

In January 2016, we announced the Proactive Safety Principles, a historic agreement between the U.S. Department of Transportation, NHTSA, and 18 automakers on a set of broad-ranging actions to help avoid the type of safety crisis that led to record-setting safety recalls. This approach is already yielding results and we're excited about its future potential.

We also embrace the lifesaving potential of automated vehicle technologies. Too often we talk about a tension between safety and innovation. A study we did showed that over 50 years, basic safety technologies such as seat belts and air bags saved 613,501 lives. Yet those technologies were also once controversial.

On September 20, 2016, the U.S. DOT and NHTSA announced the Federal Automated Vehicles Policy, a proactive, four-part measure designed to help facilitate the safe testing and deployment of automated vehicles nationwide. This approach is an unprecedented move by the Federal Government to harness the enormous safety and mobility potential of these technologies, maintaining oversight and authority, while allowing innovators to develop bold new safety and mobility solutions.

However, we can't wait until that revolution is here to address the toughest challenge in front of us: the 94 percent of crashes caused by human choice or error. We seek new strategies and new partners to help people make the right choices on the road. This fall, we convened a behavioral safety summit to lay out a roadmap to deal with issues like impaired driving, distracted and drowsy driving, and speeding.

Transportation touches every part of our lives, and the solutions to the challenges we face must likewise come from all of us. I'm confident that with NHTSA's mission-driven staff and our many dedicated stakeholders, we can make a real difference in all of our lives in the years to come. -- Mark R. Rosekind, Ph.D., Administrator

chris.pribe | 08. August 2017

@ReD eXiLe ms us,

Forgive me if this comes across a little strong, I just feel that it necessary to push back firmly against the possibility that FSD rollout would be linked in any way to V2V/V2I rollout.

Humans drive well enough with using “vehicle-to-vehicle” communications in the form of light signals (e.g. brake and turn indicators) and sounds (e.g. horns and sirens).

If some company, such as Tesla, can field a safe, self-driving car without wireless, why should we consumers have to wait for the slowest auto maker to field V2V before we can benefit?

I agree that safety is paramount. It just seems to me that regulators do best when they set performance requirements, rather than use regulation to pick winners amongst the technologies. Of course FSD technology should have some bars to hurdle, but companies insisting ahead of time that V2V is required are most likely trying to tilt the playing field.

Simply because some people promote the idea that wireless communications are necessary for FSD doesn’t make it so. If those people can’t demonstrate FSD even with wireless V2V and someone else can do so without it, who should we trust?

The fact is, the Automotive industry has been sitting on a lot of valuable spectrum for almost 20 years with DSRC. Have they delivered anything to market? They’ve made a lot of noise about Fuel Cells for even longer. Are those going to market? Can they point to any technology on the market that *they* have driven that has been remotely transformative, rather than just an iterative improvement, in our lifetimes?

Even if wireless communications can help safety (which it might, the recent collision of the USS Fitzgerald and the cargo carrier the ACX Crystal notwithstanding), is it a good idea to require V2V/V2I all around before allowing any automaker to roll out FSD (assuming their technology can clear some reasonable safety standards and actuarial metrics)?

I don’t: FSD should not be gated by V2V, because that will delay FSD for a long time. That is all I am trying to say.

eeb9 | 08. August 2017

I've worked in IT long enough to recognize some inherent risks in wireless V2V and V2I tech. Too easy to spoof and/or hack.

All it takes is one case of a V2V signal hack causing a fatality to put the skids on the whole thing.

Self-contained systems that don't rely on outside inputs are - at least for the moment - both easier to implement and less vulnerable to both accidental and deliberate interference.

That said... things like improvements in machine-readable road/lane markings and vehicle signals make perfect sense. If that's what is meant by V2V and V2I, I'm all for it. At the moment, all of those signals and markers are optimized (or, in many cases not...) for human eyes and brains. Cars can sense a much wider signal spectrum than a human eyeball, so markings and signage that can be "read" by onboard optical sensors and radar at freeway speeds are a great place to begin.

bp | 09. August 2017

First, do we expect government regulations to make technical sense?

I agree with the comment above - that human drivers use a combination of visual and audio inputs. In theory, AP2 may (eventually) have better visual object recognition than a human driver, because of the 360 degree cameras, radar and proximity sensors.

But, is AP2 including any audio input - listening for items such as approaching emergency vehicles (that may not be visible around the corner) or trains? Even with the best visual sensors, audio would still be needed to detect objects which are out of sight.

V2V/V2I might help with this, but of course there are a number of challenges to make it effective. You have to get enough vehicles and infrastructure using the technology to ensure sufficient coverage to rely on it. Then you have the technical issues related to reliable and secure communication between vehicles, which would also require enough vehicles to make the system useful.

There is a risk that Tesla is claiming AP2 has sufficient hardware to achieve FSD and receive regulatory approval. But we can't predict what the government will require in the next several years to allow FSD to be activated for humanless driving.

And like Tesla has faced with the dealership lobbying against their direct sales model, they could see similar lobbying by the other manufacturers to include new capabilities in FSD systems - which they'll be able to more easily adopt in their new cars - but will pose challenges for Tesla with years of AP2 vehicles on the road.

ReD eXiLe ms us | 13. August 2017

chris.pribe: I'm not so sure why there would be any 'push back' against what I posted at all. My understanding of the position for the NHTSA is that each fully autonomous car on the road could provide information to every other within a given region. I see it as sort of a WAZE used by the cars to be made aware of potential issues, like slowdowns due to accidents or construction or fire/life safety activity. That way other autonomous cars could navigate around the potential issue without becoming part of the same glut in traffic. I did not read it as being a necessity that vehicle to vehicle or vehicle to infrastructure communication was employed before full autonomy could be attempted.

bp: I believe the idea is that by proactively setting test procedures and fulfillment parameters to prove an autonomous vehicle can meet a certain threshold before being allowed 'In the Wild', the NHTSA is being up front, and not setting a trap for automobile manufacturers at all. There would be no surprises if the proactive, rather than reactive, regulations for autonomous vehicles are instituted properly. The government wouldn't be specifying what hardware was used, so much as which situations needed to be navigated successfully. Knowing what those are ahead of time will allow Tesla and other automobile manufacturers to tune their systems to recognize those situations and deal with them appropriately.

chris.pribe | 14. August 2017

@ReD eXiLe ms us:

I am not disagreeing with you here. You and I are of a mind about the combustion car business overall as well.

I did not aim to to push back so much as to add a ‘proactively precautionary’ note to the discussion: Mandates around technology have been known to serve other purposes than the fostering of innovation.