Autonomous Vehicle Startup PerceptIn Wants Organizations to Create Their Perfect Pod

Home > News > Content

【Summary】With $9.6 million in funding from Samsung Ventures and Matrix Partners, PerceptIn has created a $40,000 DragonFly pod that lets organizations design their ideal autonomous car.

Original Vineeth Joel Patel    Nov 06, 2018 9:30 AM PT
Autonomous Vehicle Startup PerceptIn Wants Organizations to Create Their Perfect Pod

Besides being safer than traditional cars, autonomous vehicles are expected to offer consumers with a high level of customization. Want to display a book on the windshield to read on your way to work? That will be possible, as will reconfiguring the seats to have a meeting on the go. There are a lot of things that can be done with an autonomous vehicle to make getting around more comfortable. 

PerceptIn Brings More Customization To Autonomous Cars

A new autonomous vehicle startup, PerceptIn, has come out with a truly creative offering thanks to $9.6 million in funding from Samsung Ventures and Matrix Partners that will allow organizations to create the perfect autonomous car for them. The idea is based off of the company's new $40,000 DragonFly pod. 

In a lengthy piece, Forbes got the inside scoop on the company's highly-customizable pod and it looks intriguing. Instead of offering the vehicle to the masses, the DragonFly Pod is aimed at locations that are within a closed environment: industrial parks and college campuses. 

According to PerceptIn's found and chairman, Dr. Shaoshan Liu, autonomous vehicles are just too expensive to mass produce. Liu is also wary and has some doubts over a self-driving vehicle's ability to function at safe levels at high speeds. Then there's also the public's perception on autonomous vehicles, which isn't too great at the moment. With all of that in mind, Liu believes that it's better for companies, like PerceptIn to gradually come out with self-driving technology instead of all at once. 

"This means starting with low speeds in controlled environments such as universities or industrial parks enables the industry to gain more knowledge and data so that we can reach level 4 autonomous driving in the future," Liu told Forbes. "Once people gain more experience and the industry gains more knowledge, we can move to faster speeds and eventually widespread adoption on roads." 

DragonFly Pod 3.jpg

What Allows The DragonFly To Drive On Its Own?

PerceptIn has fitted its DragonFly pod with vision-based sensors and a modular computing system that can be configured for different scenarios, claims the outlet. The purpose of doing this is to ensure that the autonomous vehicle is operating efficiently at the task at hand. So, if you're interested in only shuffling people around a college campus, swapping out the DragonFly pod's standard module for one that focuses more on human comfort versus carrying inanimate objects would be relatively simple. This level of customization is what makes PerceptIn unique in a seat filled with startups. 

"We've created a Lego-like approach and designed the vehicle so that when people are putting it together, they can reconfigure different components to use only what suits their needs," stated Liu. "We've simplified the process of building an autonomous vehicle and built a vehicle that makes the integration easy and the modular design reduces cost and the complexity of the system integration which is usually one of the most complex parts of engineering an autonomous vehicles."  

Unlike the industry standard LiDAR system, PerceptIn wants to keep things affordable, deciding to forgo the system entirely. Instead, the startup is leaning toward using radar, sonar, GPS, and computer vision more heavily. Liu claims that LiDAR is too expensive, and the doctor has a point, as the system can cost upwards of $80,000 for single use. While the startup is staying away from LiDAR at the moment, the company hasn't cut the system out of the equation for good. Liu claims that the DargonFly is so modifiable that adding LiDAR when it becomes more affordable in the future is a possibility. 

Besides making affordable autonomous vehicles, PerceptIn is also looking to become one of the few companies that helps others learn how to build autonomous cars. That goes against the grain for a lot of companies that are looking to keep secrets to themselves at all costs. 

"We've designed the building blocks to enable researchers, practitioners or even hobbyists to build AVs for their own applications," said Liu. "We are establishing joint research labs with universities and research institutes worldwide to promote ubiquitous adoption of autonomous driving technologies." 

It's a surprising move that may just work, especially as the DragonFly really does look like it has a Lego-like build to it. There's no need to dig through intricate body panels to replace parts, as they all appear to be out in the open, ready for customization. 

DragonFly Pod 1.jpg

Bringing Autonomous Mobility To Closed Environments

There's also a ride-hailing aspect to DragonFly that makes it a smart application for closed environments, especially colleges. In a video on the company's website, the company revealed the ride-hailing capabilities of the machine. Similar to what Uber and Lyft are doing, allowing users to hail autonomous vehicles in specific cities, users can also hail the DragonFly to their location.

The company doesn't provide any specifics, besides showing an individual "call" a DragonFly on a phone. The little autonomous pod then goes into action, planning the best route to get to the individual and then setting off on the course. Its systems help it detect pedestrians, bicyclists, and jaywalkers – all of which are available in spades at colleges. 

Intersections have been proving to be difficult for autonomous vehicles to tackle, as Arizona locals have expressed frustration with Waymo's autonomous vehicles being slow to act at four-way intersections. While slow, DragonFly doesn't seem to have any issues, coming up to the line, stopping, waiting for the other vehicle to pass, and then completing the turn. 

The machine than encountered, something I remember running into fondly during my college years, a stopped vehicle in the right lane. Instead of waiting to find out if the stopped vehicle was going to move, the computer saw the machine, realized that it was stopped, overtook it, and then continued on. It wasn't the smoothest process, as the vehicle zigzagged in the open lane, but it happened swiftly. 

Lastly, the video showcased a more precarious situation – one that involved both lanes being blocked by cars. In this situation, a human can remotely take control of the vehicle and get around the obstacles, ensuring that people aren't stuck waiting hours behind traffic when they need to get to class.

Some automakers and companies have spoken out against autonomous vehicles, and American drivers don't really have the best outlook on the machines. So bringing a small, useful pod to locations with individuals that respond well to technology is a good idea. And finding imaginative ways to get people around closed-off environments is a great way to get people to see the upsides of self-driving machines.

Prev                  Next
Writer's other posts
    Related Content