Like a new-born learning to use senses to make their way through life, we and getting better at equipping our vehicles with technology that allows them to understand driving conditions and the decisions needed on the road.
Data is very important to decision making, and one of the technologies used to feed data into an autonomous vehicle is LiDAR.
Essentially a laser light shot out into environment and returned to a sensor for analysis, LiDAR can be thought of as a high frequency light-based submarine sonar.
The Australian Information Industry Association (AIIA), the nation’s peak industry body for the technology sector, recently announced the prestigious National iAwards winners for 2019.
In the category of Automation Technologies Innovation of the Year, Sydney based Baraja won a merit award for their Spectrum-Scan LiDAR.
carsales.com.au chatted with Johnny Wang – Baraja’s Business Development Manager APAC – to understand what is so special about the Spectrum-Scan LiDAR technology.
Established end of 2015 Baraja has been working on the LiDAR tech for almost three years now. While consumer autonomous driving is the biggest consumer of the product, Baraja has been shipping to customers in the infrastructure, mining, heavy machinery and long-haul trucking industries
“Autonomy has been integrated into mining operations for some time now,” said Wang.
“For over a decade, what you will find with the mines is that efficiency is almost everything now, they are trying to automate other things besides the large dump trucks.
“So, the entire operation is being looked at, diggers, service vehicles and other machinery of that nature.
“The use case is so much easier than urban driving as you can imagine, no traffic lights, pedestrians, cyclists,” he added.
But we are interested in the urban driving environment, how will Spectrum Scan work to propel us on our way towards autonomy level five?
“Many types of LiDAR on the market today,” explained Wang.
“Some one hundred-plus companies working in this space, but there are very differentiating technologies that are servicing LiDAR.”
All LiDAR products need to find a way to steer the laser beam with the brute force of a mechanical spinning cylinder being the most common.
However, there are many drawbacks to this process with the cylinders being quite heavy causing vibration and temperature issues that need to be catered for. This setup can, relatively speaking, have significant breakdown issues.
Wang points out how the Spectrum Scan system can help.
“Spectrum Scan uses laser light that changes ‘colour’ often and we combine that with a prism to separate the light further into different wavelengths,” he explained.
Just so everybody knows, this is not visible light wave lengths, unfortunately Spectrum Scan does not turn your car into a travelling disco ball.
Up to four modular prisms can be mounted on a vehicle connected to a single laser light-source ‘black box’. This means there are no moving parts, and the actual bulk of the LiDAR equipment can be kept snug within a car’s body.
“With a traditional LiDAR you can see about 100 to 150 metres [ahead] with a ten per cent reflective target. Our’s is able to see 240 metres plus on a ten percent reflective target,” described Wang.
This has been combined with a high rate of analysis within the Spectrum Scan system. The result is both very high resolution recognition of the surrounding environment and the ability to discern trees from plants, dogs from people, cars, bikes and so forth.
There are, however, pitfalls to the world of cars gaining laser eyes.
“There will be a lot of LiDARs on the market in the future, and we need to ensure they do not interfere with each other,” warned Wang.
“When you send a laser beam out into free space – imagine a lot of that going on – and receivers on the other side, it’s imperative that we ensure that our LiDAR does not get interference with other LiDAR.
“Because of the way Spectrum Scan works, we are completely immune from other LiDAR and heavy sunlight,” he concluded.
This immunity comes from constant modulation of wavelengths of the egressing laser. This limits the chance of interference. However, interference will be important as more cars equipped this way, along with infrastructure to support them all start beaming into space.
This may be why Telsa’s Elon Musk doesn’t feel LiDAR is something we should rely on.
“LiDAR is a fool’s errand” said Musk during the Tesla Autonomy Day earlier this year.
“Anyone relying on LiDAR is doomed. Doomed! [They are] expensive sensors that are unnecessary. It’s like having a whole bunch of expensive appendices. Like, one appendix is bad, well now you have a whole bunch of them, it’s ridiculous, you’ll see,” he surmised.
Like just about all things in life, a little bit of lots of things might be the answer. Continue to do what Baraja is doing with LiDAR technology, but also continue to improve other sensors such as ultrasonic, GPS, optical and radar.
Keep working on AI programming and API standards will be the way forward to make the next leaps in autonomous drive.