Academy of Robotics

IMG_20191014_071507_730

A crowdcube alumni, Academy of Robotics is raising via Envestor using their platform Envestry.

Academy of Robotics are the makers of the Kar-go which is the first street legal autonomous vehicle built in collaboration with the DVLA.

The Electric green beauty was on display at the Goodwood Festival of Speed in July 2019.

Academy of Robotics plans to raise £2.5million in new investment to build more of its Kar-go.

It is to being trials on behalf of several big logistics companies in December 2019.

The Kar-go was built by an exceptional Team led by CEO and Founder William Sachiti.

The minimum investment via the Envestry platform is £25,000:
https://academyofrobotics.envestry.com/deals/2703

For investment less than £10,000 you have to contact Academy of Robotics directly via this link: https://www.academyofrobotics.co.uk/investors.html

PS: This does not must not construed as an investment advise. And for transparency, I was an investor in the crowdcube round.

At these figures this is not true crowdfunding and should be thought of as angel investing which it is.

2 Likes

It is still raising monies from a crowd even if the minimum is on a higher threshold.

I agree but at £25k that’s angel investing. I’d imagine they already have an incubator/VC invested who don’t want to muddy the cap table with 1000’s of names

The company has crowdfunded before, it’s cap table already has several hundred investors. The second link posted above is taking investments under 10K seems legit.

This is not investment advice.

They crowd funded originally from Crowdcube I have not gotten any investors update showing VC funding.

Minimum is £5,000 right now but before the Festival of Speed in July minimum was £1,000

Current state of autonomous vehicle engineering: to get the accuracy of your models to 99% in the real world you need the best of the best people in vehicle engineering, robotics (ROS, C++), computer vision (machine learning engineers/systems/scientists/etc), massive labeled datasets (good luck with that). Waymo, GM, Tesla (without a lidar), Lyft/Uber, Daimler are pouring hundreds of millions, with only Waymo (full autonomy, large datasets, testing in several cities) and Tesla (large dataset from all cars, already implemented partial autonomy on motorways) having made considerable progress. There are many well funded startups working on this in the West Coast, East Coast and Asia. Good luck to Academy of Robotics, this is very very ambitious and a highly risky investment.

Edit: in a controlled environment, such as a race track or a Goodwood Festival’s hill climb, they can have a safer working model in a lot less time. Kids can already build “donkey cars” with Raspberry Pi cameras trained to follow lanes with computer vision and reinforcement learning. On real world roads, it’s very very important that they have a massive dataset that’s constantly being updated. The risks are way higher here, as there’s no room for error with people being on the streets.

Edit 2: self-driving vehicle statups in the States seek angel and VC money.

1 Like

I invested in them when it was just a concept and to see them build everything from the ground up is a huge win. The progress they have made is phenomenal.

The start testing for logistics companies in December 2019.

It will be super challenging, they need very large funding and infrastructure, while moving fast. Perhaps they are hoping to partner with a car manufacturer.

Overall, I think it’s exciting that UK/Europe is trying to keep up with Waymo and other “startups” which already have a few years of datasets and experience. Lyft is hiring people in London for its Level 5 project.

Nissan, which has been testing for a while, is also nowhere near and they have deep pockets to hire and map the areas.

A full list of UK robo-car projects can be found here:

To understand the tech a bit better, below is a June 2019 presentation from a research conf. workshop by a company set up by ex-Waymo (Google), Tesla Autopilot and Carnegie Mellon University/Uber Self Driving Car project leads:

They have had offers which were in every sense very good but the caveat is the production will move East. They would rather keep production in Europe for as long as it possible.

Among machine learning/AI researchers, there’s a general understanding that the hype is real but the tech will get better and more useful in a few years, just not yet. It’s really crap right now with lots of compute power finding useful and useless relationships in vectorised data (text, signals, images, sounds, etc).

With the current tech stack — new hardware and feature engineering on top of old methods (neural networks) — having self-driving cars with 99.99% accuracy is still out of reach, giants like Waymo says Level 5 it may never exist.

John Krafcik, head of the self-driving car unit of Google parent company Alphabet, said that though driverless cars are “truly here,” they aren’t ubiquitous yet. And he doesn’t think the industry will ever be able to drive at any time of year in any weather and any condition, the highest driving rating. Driving in all conditions can be difficult for humans too, he noted.

“Autonomy always will have some constraints,” he said.

“It’s really, really hard,” Krafcik said. “You don’t know what you don’t know until you’re actually in there and trying to do things.”

… Waiting for better universal function approximators from those PhDs.