Shadow Driving will Ruin the Autonomous Industry
The overall problem here is that the use of public shadow driving can never lead to autonomy. And much worse it will cause thousands of avoidable injuries and deaths and decimate many of the vendors in the space and set progress back decades. The “conventional wisdom” and echo chamber are wrong on this. (Keep in mind the vast majority of those with the “conventional wisdom” has NEVER made an autonomous vehicle. They have no real experience with the technology like NASA and aerospace does and has had for decades). Can someone show me data from anyone who can refute the reasons why this will occur below? From someone who has actual experience with an end state product? Not one where they are still in early stages of development driving easy or benign scenarios? Can you give me reasons to not switch to aerospace level simulation for AI, engineering and test?
One Trillion miles need driven at a cost of over $300B — supported by Toyota and RAND who states it is impossible.
Shadow Driving is Dangerous. People will not pay proper attention, especially for long periods of time, no matter what you do. Supported by NASA, Clemson, University of South Hampton, Volvo, Ford, Waymo and Chis Urmson. That will lead to thousands of avoidable injuries and causalities when the develops move from benign scenarios to the complex, dangerous and crash scenarios. Which have to be run thousands of time each.
The tide is clearly shifting. It can be seen in the rising amount of professionals on LinkedIn and in press articles coming out week to week. But it’s not shifting fast enough. We need to avoid the first death of a child or family. What we need it for what is for the majority remaining to put their ego aside, look at the data, realize there are other industries far more experienced than you who have something to offer and help change from shadow driving to simulation. Data and common sense are not on your side. If you persist that first child or family will perish needlessly in an L2+/L3 vehicle. When that happens, and it has to by your plan, the public, government officials, the press and lawyers will finally get it. They will see you were wrong the whole time and doubt your competence or ethical fortitude. That will bring far more regulation and a far longer delay than if we rise to the occasion now and self-police.
This recent video of a Tesla driver saving this autopilot system if the very tip of the iceberg and nearly 100% avoidable.
What happens when it is a more complex and dangerous scenario and a family is on board? When the weather is bad, the road is bad, the scenario is worse and there is a vehicle to the left of where this person bailed to? And now run thousands of scenarios like that thousands of times. You will have thousands of accidents, injuries and casualties.
All of the folks using shadow driving and saying or believing they are creating this technology to save lives are causing far more harm than good. The benign scenarios you are running now is masking that. The belief that shadow driving is the best or only way to go about this and that the thousands of injuries and deaths that are coming are necessary to save many more lives later is WRONG. That ends does not justify the means. You are creating a far worse problem than you are trying to avoid. You will, by your plan, harm thousands of people needlessly and set the industry back decades. And probably ruin your company and be driven out of business. Please do the industry and those who will benefit from the switch to simulation a favor. Do you due diligence, think outside the echo chamber, deal with your ego and help us make the switch. Or show us where we are wrong.
Update 8–23–2017
Waymo has now switched to simulation from shadow driving and is using MCity’s process to create core scenarios. Just a short time ago Waymo announced Level 3 should be skipped. Waymo is doing this for a reason. Others should pay attention.
More details on the issues and solution can be found in these articles
Autonomous Levels 4 and 5 will never be reached without Simulation vs Public Shadow Driving for AI
https://www.linkedin.com/pulse/autonomous-levels-4-5-never-reached-without-michael-dekort
Who will get to Autonomous Level 5 First and Why.
https://www.linkedin.com/pulse/who-get-autonomous-level-5-first-why-michael-dekort
Due Diligence Recommendations for the Mobile, Autonomous and Driverless Industry
https://www.linkedin.com/pulse/due-diligence-recommendations-mobile-autonomous-industry-dekort
