Advanced automated driving being tested using an F-150

[Aron]

Here's an article discussing how researchers are using an F150 to test out computer-assisted driving that can react to unusual situations (debris in roadway, work zones, emergency vehicles, etc.). I find it interesting that they seem to be adding cameras and other sensors (that looks like a gps antenna in the photo), rather than relying upon the F150's native gear. The projected registration screen for law enforcement is a pretty cool idea as well.

https://www.washingtonpost.com/transportation/2023/11/02/self-driving-testing-dc-395/

Sponsored

 

[ReverendQ]

Here's an article discussing how researchers are using an F150 to test out computer-assisted driving that can react to unusual situations (debris in roadway, work zones, emergency vehicles, etc.). I find it interesting that they seem to be adding cameras and other sensors (that looks like a gps antenna in the photo), rather than relying upon the F150's native gear. The projected registration screen for law enforcement is a pretty cool idea as well.

https://www.washingtonpost.com/transportation/2023/11/02/self-driving-testing-dc-395/

It doesn't appear they used an ankle weight.

 

[HammaMan]

They're tapping the IPMA connectors like the comma device is doing. I hope to see as things progress that we can swap in a 4D radar as it's capable of picking up road debris among other things.

This approach being used here should be able to control any sync 4 vehicle, but the additional sensors being added on just isn't the way to do it. Upgrade units on the vehicle.

 

[superfordman]

For us non paywallers, https://news.vt.edu/articles/2023/10/research-autonomousvehicledemo.html

I just wonder how this new feature found in the 24 Mustangs will impact things, as I bet money it'll be in the 24 F150 if not the following year(s).

 

[HammaMan]

I don't think people understand where tesla's public beta is on FSD using vision alone. All of the ancillary stuff mentioned is easy add. Actually navigating difficult areas is where the true costs add up. Trashing all of their previous software a year ago has led to better results in a small fraction of the time. All of the additional sensors and cameras are truly redundant. Tesla hasn't even begun to leverage their "4D radar" (that's what the tech is called) which is an amazing piece of kit.

The conditions seen here are easily some of the most challenging out there, hands / feet free. This is the older version and the newer tops it but it's not open public as of yet but small clips can be found (it's trained by the "best" drivers as determined by tesla)



As for encryption on the canbus, there's sensor hijacking that's as old as time, there's also the possibility that strict encryption policies might not be in place presenting a vector for exploitation or a MITM attack via a new device install handshake. I wouldn't expect to see a full implementation until ford in-houses everything, after they get a competent dev team fully up to the task that encompasses all of those aspects. Thus far they've been rather unimpressive give just how much $ ford has to do it all properly. They've also got former tesla guys that you know are pulling with all of their might to right the company for what consumers demand -- but it's bloated legacy auto which is a big ship with a tiny rudder. They have to trim as much of the old guard as possible in order to effectively be able to pivot the company as market demands.

Sponsored