Atlanta's Chance to Rise Again - This Time Multi-Modal

Warning: Not on driverless topic.

For the whole night after the Atlanta interstate implosion - and the giant fire which accompanied it - all a quick walk from my daughter's apartment, I thanked the divine deity (or deities), or the universe, in the event that the atheists are correct, for the miracle that my daughter and no one else in Atlanta was hurt. Absolutely amazing that on this usually congested stretch of highway that no one was killed or severely injured.

But now I am up to the next phase, the transportation solutions phase, of figuring out how metro Atlanta can cope beyond the advice this morning of the desperate Georgia DOT secretary, who advised everyone to telecommute. That is a wonderful partial solution and for the first few days it might even be the entire solution and perfectly possible for many commuters. However, telecommuting is a piece of the pie and certainly not the whole pie as this mess goes moves into its second week or then into a months-long reconstruction project.

From the baseline

This is Atlanta's - and Georgia's - golden opportunity to go multi-modal. Yes, there is a little subway system, but it will be filled to the gills. Yes, there are buses, but their poor frequency, reliability, and crowds will not suit them to be more than a small piece of the solution to the commute clusterfu*&k that we will soon see.

Now is the time for Atlanta to shine.

Sun shines, we all hold hands, sing Kumbaya 

Here is how, in my own totally non-humble, confident opinion, the city of Atlanta and the Georgia DOT should proceed to reach a happy ending. Remember the magic two words: ALL TOGETHER. No one strategy will be the solution.

1.  Impose an impromptu HOT lane, to promote slugging - otherwise know as casual carpooling. If you charge an few dollars, maybe even one dollar, for regular lane use, then people will be happy to catch rides at designated, convenient places. This happens safely every single workday in Virginia, Washington, DC, San Francisco, and suburbs over the Golden Gate Bridge.

2.  Put down paint and to open up lanes for biking on major roads. Putting in cones or poles will help as well. Imagine a biking highway or major regular thoroughfares inviting bikes so that the roadways can fit more people in a smaller space. This is cheap, both for commuters and for local and state governments.

3. Put on the road lots of cheap or – better – free express buses – I’m dreaming of dedicated lanes here, but I will settle for less – on significant commuter routes. One hundred buses can move 6000 people. That is many more people, moving much more efficiently than in SOV cars.

4. While we’re at it, make some improvements, tweaks really, to traffic signals and sidewalks so that transit and other modes are workable for those pedestrian connections that are necessary in a world beyond strolling through office and strip mall parking lots.

5. Telecommuting.

6. Apps, apps, app. Let all the data be open so that enterprising people can develop imaginative solutions, some of which will take off and really help with the challenging situation.

7. Invite taxi and ride hailing (aka TNC) companies to offer more attractive shared-ride services. 

Needed: Middle-of-the-night painters/civil disobedience and/OR wise governance

None of these projects would be expensive. But together, - remember the magic two words, ALL TOGETHER – all cheap and quick fixes – can demonstrate to metro Atlanta what it has hitherto (don't get much chance to use that word) refused to admit – that it can become a magical region with 20-minute neighborhoods and easy, reliable, attractive options for connectivity. Goodbye awful auto-centric congestion. Hello Brooklyn, I mean Portland, I mean the rising of a new Atlanta and the best of a new South. 

Better for this episode to be remembered as the rising of a new Atlanta rather than yet another burning of the Peachy-keen city.

Legislation Bonanza - North Dakota Paragraph

The North Dakota bill, HB 1202, is one paragraph long and it only mandates that a study be performed. It's a quick read.

Here it is, with my comments inserted.

SECTION 1. DEPARTMENT OF TRANSPORTATION STUDY - REPORT TO LEGISLATIVE ASSEMBLY. The department of transportation, in collaboration and consultation with the autonomous vehicle technology industry [Obnoxious editor's note: You do not need "the ... industry" to study this], shall study the use of vehicles equipped with automated driving systems on the highways in this state and the data or information stored or gathered by the use of those vehicles [Obnoxious editor's note: A previous North Dakota bill was concerned about the privacy and cybersecurity of driverless-generated data]. The study must include a review of current laws dealing with licensing, registration, insurance, data ownership and use, and inspection and how they should apply to vehicles equipped with automated driving systems. The department of transportation shall report its findings and recommendations, together with any legislation required to implement the recommendations, to the sixty-sixth legislative assembly."

Legislation Bonanza - The Peach State

Georgia, the state where everything is named either Peach or Peachtree, is considering a few bills. (Okay, my knowledge of the state is minimal. I have only been to Savannah and Atlanta. I still have not visited the Coca Cola Museum.)

Senate bill 219

This bill has already been passed by the Georgia House of Representatives. The legislation definitely conceives of and includes provisions for the routine operation of fully autonomous vehicles.  The bill would also regulate "electric assisted bikes" and "electric personal assistive mobility devices," which, if I am reading correctly, are power scooter wheelchairs. 

This bill expressly states that there is no need for any licensed humans to be present in a vehicle. Indeed, the vehicle may be operated while empty. Autonomous vehicles are added to the list of exemptions from the Georgia drivers license requirement.

Senate bill 218

According to SB 218, if a human driver is present in the vehicle, that person is deemed the operator. However, the bill allows for driverless vehicles to be operated without a human on board, in which case the technology is deemed the driver or operator. 

Some kind of black box technology is required to be on each driverless vehicle so that collision details will be recorded.  

The bill specifically states that NHTSA regulations will supersede state laws regarding autonomous vehicles. The legislation authorizes the Georgia Department of Public Safety to carry out the law, if passed, of course.

Senate bill 54  

Peach State political representatives are apparently divided on the driverless issue. This bill requires that there be a human driver in the vehicle - unless it is being driven on a closed course - and that there is a means for said human to easily engage or disengage the autonomous mode. 

So much for a smooth peach puree path to a completely driverless future.

The bill mainly confers on the state department of transportation the responsibility for filling in the details, but this legislation only provides for testing of driverless vehicles on Georgia roads and not for the routine operation or sale of such vehicles.

Since I am blissfully unaware of the details of Georgia politics and its state legislature, I will avoid speculation. 

Legislation Bonanza - Focus on Pennsylvania

Would everyone just stop already? I do not have time for all things driverless at this rate. TMI - too much information. And one must actually read the information to determine whether it is (a) info vomit, (b) mildly significant, (c) or truly valuable.

But that is life, no?

Pennsylvania legislation

Pennsylvania (PA) Senate bill 427 - A long, detailed bill, presumably designed to sidetrack the state DOT (my opinion) in making detailed decisions about driverless testing and applications to test autonomous vehicles. The legislation does not concern sale and normal operations of AV in everyday life on state roads. 

This is not only an AV piece of legislation. The bill makes a big deal of addressing platooning and offering analogous provisions for platooning at every opportunity. That contributes to the length of the bill, obviously.

Pretty meaningless department

(1) An advisory committee of mostly state officials who are required to meet at least once a year to discuss AV and platooning issues. Insert - sigh here; harrumphs are acceptable.
(2) Report due by Jan. 1, 2020 that is supposed to cover the AV kitchen sink of issues.
(3) Another report due by Jan. 1, 2020 to address barriers to AV and platooning testing, deployment, and operation. 
(4) A third report due by a different date, this time Jan. 1, 2025, to address infrastructure improvements and funding relating to AV and platooning. This legislation will definitely have been superseded by then. 

Standard provisions

Ability to satisfy a judgment for damages.
Preemption of local laws.

Interesting stuff

Special sticker on every AV or platooning vehicle being tested. No word on size or design of the required sticker. (My obnoxiousness is visualizing the you're-great-stickers that pediatricians give to kids at the end of each appointment, the one that even four year olds don't care about.) (One more parenthetical: Uber already does the sticker thing.)

Approval or denial of applications to test AVs must be issued within 15 business days, but the public will be entitled to comment. A quick process; I question how meaningful or promoted the public comment right will be.

Level 4 and 5 vehicles being tested may be controlled remotely. There is NO requirement that a licensed human be on board and ready to take over the wheel.

There is a provision that mandates reporting of cybersecurity breaches. 

Instead of using either the word "crash" or "collision," the bill employs the word "accident." After an accident with a vehicle being tested and the requisite reporting of the episode, the bill would allow for "self-certification" explaining how the tester will "reinstate" the AV or platooning vehicle back into service.

An audit once every two years of every permit. The process will also be designed to examine safe testing and management of AV and platooning. 

More stuff to read ...

New Kids on the Block

You would think the autonomous vehicle world would wait a respectable time after my relaxing vacation to spawn new competitors and fill my inbox with lots of news. But no, just this week I am reading about deals, laws, and now two start ups that threaten to do better what is already being attempted. If these start ups meet with success, there is little doubt that billions - with a B - will be offered to acquire their talents.

Instant AI 

DeepScale - An AI company that has burst on the scene with "deep neural nets and computer vision systems." This is something Nvidia is involved with, but this new company claims to much more quickly and efficiently teach the software to navigate road situations. DeepScale is working on the perception capability of driverless vehicles. Thus far, the company has attracted $3 million - with a mere m - in funding. It's chicken feed by current standards, but it gives hope and some life to about 20 people who are way smarter than me.

Do not confuse DeepScale with Deep State, which I learned about from Samantha Bee. It is the far-crazy-right-wing (yes, even some in the Administration's undocumented theory - why trouble a good theory with facts?) that government workers and Obama are hijacking the federal government.

Take two cameras and transform the jalopy

AutoX founder - former assistant Princeton professor Jianxiwong Xiao - bought a few cameras at Best Buy, added some AI software, and presto jammo, he might be transforming the nascent driverless vehicle industry. No GPS here, but he expects to add LIDAR cameras down the road. Xiao has left Princeton and moved to California to pursue the dream.



Who would not want to move from even a pretty part of New Jersey? I have to admit, though, I would pause at a move to Silicon Valley. I do agree with the avoidance of GPS. Not a fan.

Xiao says it will likely be two to three years before an AutoX system is ready to be sold to vehicle manufacturers. Yes, here's the video again.

Okay, now on to reports and lots of legislation at the US state level in the works.

Nividia Versus Mobileye - and They Are Not Alone

Nvidia is confident. It has a buddy in Tesla owner Elon Musk, it has truck stuff going on with PACCAR, a major truck manufacturer, and Wall Street is enamored.

More than apples

They do more than grow apples, launch ships, and grow computer companies out there in the other Washington - Washington State that is. Now they are happy as clams in their expectation that truck manufacturer PACCAR's partnership with Nvidia will bring lots of driverless truck manufacturing to the Evergreen State. Others, probably in states with more professional truck drivers, are concerned about the loss of trucker jobs and jobs in the industries that serve those drivers. Waffle House? Flying J? Where will they be without the middle class disposable income of truck drivers?

Cheap can be good

Forget expensive LIDAR and sensors, this Dr. Jianxiong Xiao, a computer vision professor at Princeton,is using $50 cameras from Best Buy, a US chain store that sells TVs, laptops and batteries. It's not a high-tech mecca. The company is AutoX and an article describes the non-Mobileye approach to driverless.

Yes, the video of the self-driving AutoX car is here.

Let's face it, there are reports each day about up and comers in the software engineering, cameras, and LIDAR corners of the driverless supply chain. There are still plenty of start ups willing to push their ideas and compete with the big guys in this still fluid market.

And from somewhere in the Mediterranean ...

Mobileye has money now that Intel has purchased the company. The question is whether Mobileye can maneuver past Nvidia with its over $15 billion infusion of cash from Intel. Gee, no one talks millions anymore. However, I am not the only one doubting Intel's strategy. Is Intel grasping to have a presence in a market it is unfamiliar with? Did it overpay and go the wrong route as far as artificial intelligence?