Skip to main content

One Crushed Tesla, One Big Dispute


Autonomous features are one of Tesla’s big selling points, even if they aren’t all part of the production software for the firm’s cars. One such feature is Summon, which is designed to allow a driver to park in a narrower space than would normally be possible, by autonomously entering and exiting a parking space with the owner outside the car.

As it’s in Beta, owners have to sign up to enable the feature and are given specific instructions on how and where it can be used. Even so, one owner has ended up with a rather crushed EV and the resultant dispute over who is to blame isn’t going to do Tesla’s reputation much good, whether it’s in the right or the wrong.

The issue here is whether Tesla’s implementation of the self-parking feature fails safe. Of particular importance when it is designed to be used without a driver in the car.

The owner claims that he hadn’t initiated the Summon request, Tesla disagrees. That raises all sorts of issues. Tesla’s logs of activity, taken from the vehicle itself, suggest that all of the requirements for the self-parking feature were met by the owner. And they do seem pretty comprehensive.

Far from conclusive though.

The problem seems to me that Tesla’s process for initiating Summon is too close to the normal process of exiting a vehicle, close enough that an owner could inadvertently enable the feature. The process of double tapping the gear selector, selecting park and releasing the brake pedal is 95% of what every automatic transmission vehicle owner does when exiting their vehicle. The double press of the gear selector stalk seems like something that could be done entirely accidentally.

Rather than getting into a dispute with one of its customers, Tesla needs to suck it up and accept that, in this case the logs are irrelevant.

A self driving feature must be both fool proof and fail safe. In the case of Summon that patently isn’t true. Fortunately the only damage here is to the company’s reputation, the owner’s ego and of course the sheet metal itself.

Fixing the feature seems easy to me. If the driver isn’t in the car Tesla needs to enable some form of secondary control to ensure the vehicle is doing exactly what the owner wants it to do. That could be a button on the keyfob that needs to be pressed throughout the self-parking manoeuvre or a control in the car’s smartphone app.

Autonomy is an extremely difficult feature to enable for something as potentially dangerous as a motor vehicle. This strikes me as the first of very, very many disputes that will arise as the technology broadens and matures.


Popular posts from this blog

F1: Robert Kubica Impresses In Renault Test Run

The car may be old but its the performance of the driver that's the story here. Robert Kubica returned to F1, after a fashion, earlier this week with an extensive test run in a 2012 Lotus Renault F1 car at Valencia.
The age of the car and the circuit were likely determined by F1's current rules which ban testing, but the reason for Kubica being in the car is far more interesting. Considered by many to be a potential World Champion and certainly one of the fastest drivers of his generation, Kubica's F1 career seemed to be over after a 2011 crash whilst driving in the Rally of Andora. His Skoda Fabia was penetrated by a guardrail in the high speed accident partially severing his right arm.
Up until last year Kubica has been competing in rallying, with the expectation that the limited movement in his repaired arm would prohibit a return to single seater racing.
So this week's test is both interesting and confusing. Interesting because Kubica completed 115 laps of the ret…

Panos Panay's Defence Of Microsoft Surface Hardware Sounds Eerily Familiar

This weekend I went out with my ten year old daughter to select a laptop for her school year beginning in January. The schools requirements are quite specific, requiring a Windows 10 device, with a preference for a touchscreen and a stylus. She chose a Surface Pro, after trying a large number of different options. Having seen the way I use my own Surface Pro - and tried it herself there was only ever going to be two options - and the other was a Surface Laptop.
I tell you this so that you understand I am a buyer of Microsoft's products through choice, not compulsion. I'm on my third Surface device now. 
So when Panos Panay dismissed reports of the death of the Surface hardware line, I was very interested to see exactly how strong these denials were. Especially how they reflect what has gone before. To whit: Windows 10 Mobile.
Panay claimed that Microsoft is in hardware for the long haul. Almost exactly mirroring the words of Terry Myerson, when he claimed Windows Mobile was g…

WhartonBrooks Indiegogo Windows 10 Mobile Even More Doomed To Failure Than Usual

WhartonBrooks is currently crowd-funding its latest Windows Mobile smartphone on Indiegogo. If crowdfunding isn't already a bad enough idea, a company trying to crowdfund a Windows Mobile device should be warning enough for you.
Not that anyone seems to be taking the project too seriously. With a few weeks left to run the campaign has managed to ensnare just 2% of its $1.1m target.
If you want a better indication of how few Window Mobile loyalists remain I doubt there is one. Of 3,900 Windows Phone enthusiasts Wharton Brooks was seeking for its new phone, it has managed to entice just 50.
Windows for Phones is dead, even if the corpse hasn't stopped twitching yet.