Better safety through radar

jerhenderson

Well-known member
First Name
Jeremy
Joined
Feb 20, 2020
Threads
4
Messages
957
Reaction score
1,415
Location
Prince George BC
Vehicles
Cybertruck
Occupation
Correctional Officer
Country flag
Safety is what I want most from my Cyber Truck. (I want all the other great features too.) Therefore I want it to see what I can’t see with my eyes that could keep me safe when human vision can’t detect what is happening. The Cyber Truck should see in/through the fog, through the spray of water that comes up from a semi or other vehicles on the rainy crapy roads and freeways of the pacific northwest. I want it to see through the snow flying up from behind another vehicle so common in winter in cold climates like Montana etc. In order to do this it needs some kind of active vision like radar or lidar. If the FSD only works in fair weather I can do that, I want my amazing machine to do what I can’t do. To rely only on regular cameras is limiting. Unless Tesla can tell me otherwise that their camera can see through fog etc I want it to have radar or something to see what human vision can’t.
guess your s*** outta luck then.

 

ajdelange

Well-known member
First Name
A. J.
Joined
Dec 8, 2019
Threads
4
Messages
3,214
Reaction score
3,363
Location
Virginia/Quebec
Vehicles
Tesla X LR+, Lexus SUV, Toyota SR5, Toyota Landcruiser
Occupation
EE (Retired)
Country flag
Many more photons make it through a fog or a dark night than we can observe with our eyesight.
If you are thinking like Hercule Poirot, IOW, distinguishing between seeing and observing that is correct. We see all the photons but an image is made by deconvolving the received image with the sensor transfer function. Our brains have deconvolution filters that are deigned tp give ius the information we need at the rate we need it in the format we need it. We cannot deconvolve an image that we see through fog. That's what the California scientists are trying to do.

In fact, Lidar is exactly this. The laser is shot in a scanning motion and only covers all points in the field over a period of time. You have to integrate all the returns over this period to form an image of what is in front of you.
The problem with LIDAR is that you don't see what happened in a particular location when the beam was pointed elsewhere. LIDAR is sparse in time and space. What Musk is talking about is sampling every pixel at a high rate simultaneously using cameras that don't scam. Thus we have a high sampling rate in 4 dimensions and, if only we can deconvolve, we can get a map of 4-space. Of course that is no trivial task.

Lately I've been working with nadir-pointing helicopter-flown aerial lidar data. The lidar is capable of seeing through the forest canopy and mapping the ground surface below.
I was introduced to this in 1967, I think it was, when it was called "gated TV" or something like. It was used to observe the VC scuttling about beneath the canopy during the Vietnam fiasco.
 

CyberGus

Well-known member
First Name
Gus
Joined
May 22, 2021
Threads
17
Messages
1,315
Reaction score
4,040
Location
Austin, TX
Vehicles
1981 DeLorean, 2023 Cybertruck
Occupation
IT Specialist
Country flag


rr6013

Well-known member
First Name
Rex
Joined
Apr 22, 2020
Threads
43
Messages
1,242
Reaction score
1,197
Location
Coronado Bay Panama
Website
shorttakes.substack.com
Vehicles
1997 Tahoe 2 door 4x4
Occupation
Retired software developer and heavy commercial design builder
Country flag

rr6013

Well-known member
First Name
Rex
Joined
Apr 22, 2020
Threads
43
Messages
1,242
Reaction score
1,197
Location
Coronado Bay Panama
Website
shorttakes.substack.com
Vehicles
1997 Tahoe 2 door 4x4
Occupation
Retired software developer and heavy commercial design builder
Country flag
Eltric Work&Fun said:
I want my amazing machine to do what I can’t do. To rely only on regular cameras is limiting.
Tesla BEV‘s using EAP and AP which have statistically exceeded human driving reported accidents IIRC.

Detrimental reliance upon vision(PUREvision) in Tesla Autopilot is beyond the limits of human eyesight. Regular cameras Tesla uses in rearview/backup and interior cabin monitoring.

Tesla recently sign contract with Samsung to supply its (greater than human) high resolution photometric image sensor behind Tesla camera lenses. This is the reputedly ToF(Time of Flight) measure, invisible to the naked eye, of “photons” returned to its sensor. SO a higher level of performance exceeding human eyesight Tesla is using in HDWE4.0 and included in the Cybertruck.

ToF enables 4D which in technical terms provides Teslas additional information humans equate as “range finding”. Its uses are inside the HDWE stack are realtime, quicker than wet hardware. And the benefits accrue to improvement in Safety upping precision at speed.

Edge cases and corner cases apply wrt: limits. Meat bags 70% water that we are defines the absolute limitations however much technology approaches, extends and excels human.
 

ajdelange

Well-known member
First Name
A. J.
Joined
Dec 8, 2019
Threads
4
Messages
3,214
Reaction score
3,363
Location
Virginia/Quebec
Vehicles
Tesla X LR+, Lexus SUV, Toyota SR5, Toyota Landcruiser
Occupation
EE (Retired)
Country flag
adds noise that contributes to more processing filtering.
If the GDOP goes down more than the noise goes up the estimate is improved. That's the reason one has multiple sensors. Telsa's argument for dropping radar was that the overall gain improvement from it was not worth the extra processing effort..
 

CyberGus

Well-known member
First Name
Gus
Joined
May 22, 2021
Threads
17
Messages
1,315
Reaction score
4,040
Location
Austin, TX
Vehicles
1981 DeLorean, 2023 Cybertruck
Occupation
IT Specialist
Country flag
Ignoring newstand pop psychology ref., agreed danger is not risk nor risk assessment. What’s so telling about that and RADAR?

Is the intimation that Tesla is using <xyz> to assuage both to further FSD?
I'm suggesting that an autonomous vehicle driving exactly like a human would be deemed unsafe, even though millions of humans drive that way daily.

NHTSA forced Tesla to recall FSD beta due to its imitation of human "slow roll" though stop-signs, even though this caused no accident nor injury.

https://insideevs.com/news/564570/tesla-relcall-fsd-rolling-stops/
 

Cybertruckee

Well-known member
First Name
Yosemite Sam
Joined
Apr 21, 2020
Threads
7
Messages
331
Reaction score
371
Location
Mostly under the pines
Vehicles
Red Rider
Occupation
Finance Manager
Country flag
My son who works on aviation technology says that radar now is so advanced that it can tell whether a pile of white stuff on say a surface of runway is a pile of snow or concrete barrier.
 


Ogre

Well-known member
First Name
Dennis
Joined
Jul 3, 2021
Threads
107
Messages
6,532
Reaction score
15,369
Location
Ogregon
Vehicles
Model Y
Country flag
I’ve seen little evidence that adding Radar would fix the issues which FSD has at the moment.

There is just this blanket assumption that “moar” sensors is better. With this assumption you end up with a vehicle with Radar, Sonar, LiDAR, FLIR, and vision.

Adding something makes sense if it is solving a specific problem. I don’t see that being the case here.

Look at the videos of FSD in action, from what I’ve seen the big problems are not things like figuring out how far away something is. The big problems I’ve seen with FSD involve figuring what is a lane, making unprotected left turns, where the lanes go at intersections and things like that. Little of which Radar would help with at all.
 

ajdelange

Well-known member
First Name
A. J.
Joined
Dec 8, 2019
Threads
4
Messages
3,214
Reaction score
3,363
Location
Virginia/Quebec
Vehicles
Tesla X LR+, Lexus SUV, Toyota SR5, Toyota Landcruiser
Occupation
EE (Retired)
Country flag
There is just this blanket assumption that “moar” sensors is better.
That's actually a good assumption.

With this assumption you end up with a vehicle with Radar, Sonar, LiDAR, FLIR, and vision.
Yes, you would were it not for the fact that covariance of the parameter estimates is not the only thing under consideratiion here with cost, marginal improvement added by a sensor, complexity, weight, power consumption, component availability.... being some of the others.

Adding something makes sense if it is solving a specific problem. I don’t see that being the case here.
You don't have the relevant experience, perspective nor knowledge of this system to offer an opinion (and I don't either). The people who do took radar out for some of the reasons I listed above. They didn't think what it added justified the cost of developing the software to integrate it properly. Now there are some indications they are putting it back in. We will see.
 

FutureBoy

Well-known member
First Name
Reginald
Joined
Oct 1, 2020
Threads
185
Messages
2,072
Reaction score
3,432
Location
Kirkland WA USA
Vehicles
Toyota Sienna
Occupation
Financial Advisor
Country flag
The people who do took radar out for some of the reasons I listed above. They didn't think what it added justified the cost of developing the software to integrate it properly. Now there are some indications they are putting it back in. We will see.
From what I've read, the radar that might be getting added back in is sufficiently different enough from the original radar that was taken out that I'd classify them as really 2 different types of sensors. But yes, at a high level, radar was removed from the system and now there are indications that radar is possibly getting added back in.
 

CyberGus

Well-known member
First Name
Gus
Joined
May 22, 2021
Threads
17
Messages
1,315
Reaction score
4,040
Location
Austin, TX
Vehicles
1981 DeLorean, 2023 Cybertruck
Occupation
IT Specialist
Country flag
My son who works on aviation technology says that radar now is so advanced that it can tell whether a pile of white stuff on say a surface of runway is a pile of snow or concrete barrier.
Sensors with such high resolution are cost-prohibitive for automobiles.
 

Crissa

Well-known member
First Name
Crissa
Joined
Jul 8, 2020
Threads
86
Messages
9,599
Reaction score
15,195
Location
Santa Cruz
Vehicles
2014 Zero S, 2013 Mazda 3
Country flag
My son who works on aviation technology says that radar now is so advanced that it can tell whether a pile of white stuff on say a surface of runway is a pile of snow or concrete barrier.
Radar has always been able to do that. Because everything is some amount of transparent to radar.

The problem is that it's not so good at telling exactly where things are, it sees echos and harmonics everywhere.

It's terrible at seeing textures on surfaces, or telling whether something is over the horizon or on the side of the road. What it can't tell if something is a parked car or just a wall on the side of the road.

It's a problem of resolution and processing power.

-Crissa

 

 
CYBERBACKPACK
Top