How far are we from an integrated LiDAR eco-system?
NSSN’s Data Research Theme Leader, Dr Zhitao Xiong, shares his thoughts on the future of LIDAR. The need for LiDAR (Light Detection and Ranging) sensors in the autonomous vehicle industry or related Advanced Driver Assistance Systems (ADAS) cannot be overstated, because of its capability in establishing 3D representation of objects in real-time. It can help the autonomous vehicles as well as ADAS to detect surrounding objects for their positions and dimensions more accurately than other sensors such as cameras and radars.
Grown up from its early form LMS (Laser Measurement Systems), LiDAR has experienced a dramatic growth in recent years in the automobile industry, although it has been used in surveying for many years. I still remember the old days when autonomous vehicles were still called Autonomous Land Vehicles (ALV) and when I was cautious in fixing a SICK LMS 200 sensor onto it – the sensor was so expensive that the team agreed damaging any part of it would lead to the end of our master study. If my memory serves, this was back in the year 2007.
How about now? A current LiDAR model could cost you only $600 or if you are rich enough, you could get a $75,000- one. Of course, to build a 3D representation of the surrounding environment, one LiDAR sensor is not enough as it only covers a certain scanning angle, e.g., 145⸰ , not 360⸰
So, LiDAR is affordable and not particularly out of reach for us at all, isn’t it? The good news is, car manufacturers such as Toyota and Subaru are starting to include cameras as part of their ADAS for the mass consumer market. Cameras are here, how far is its friend LiDAR? Well, not that far, but not close either. It will be a longer journey than we thought, as it will come all the way through a hardware, software and business adventure, the end, is a LiDAR eco-system that could eventually support the mass market as shown below.
The market is there and you can turn this into your own dollar farm. We have already seen a start-up community in the LiDAR sector grow in recent years - investors observed such farming opportunity in supporting autonomous vehicles and ADAS mass deployment, which resulted in at least 64 LiDAR-related start-ups in the USA, 31 in Europe, 6 in China and 6 in Australia. In NSW in particular, we have seen the emergence of Baraja and Ocular Robotics.
Then how to build your own eco-system to survive this fierce competition? A competition of dominating a market of your own as well as securing external growth funding? This is how NSSN could help using our “NSSN’s guide to the LiDAR Adventure”. As the catalyst for smart sensing technologies with seven member universities across New South Wales, NSSN is working hard to integrate and cultivate the smart sensing eco-system by connecting industry, academia and agencies. We are passionate in working with clients on technical solutions as well as novel business models, drawing upon the research expertise of some of the brightest minds in the state. Then what is NSSN’s guide to the LiDAR Adventure?
If we start with the technology part sitting on the left, we could be busy in identifying acronyms and specifications: beam-steering, distance measurement method, wavelength, scanning angle, range, etc. However, no matter how point cloud is collected, the following three things would matter to the end users most:
1. Performance: this could indicate the range of the LiDAR, scanning angle of LiDAR, power consumption and safety. Remember, LiDAR is still a laser-based system, it is possible for a LiDAR system to damage a camera for instance;
2. Dimension: this could indicate how LiDAR sensors are going to be installed in the vehicles without distracting the attentions of car users or pedestrians and integrate perfectly with the overall safety design of the vehicle;
3. Price: this would indicate the acceptance of the LiDAR in the market. This should be a balance among performance, dimensions (fit for specific design purpose of a vehicle) and the profit that your company would like to gain.
But, the output of the technology will be ultimately used in driving decision making algorithms, such as to locate the autonomous vehicle of interest in a junction or to avoid a pedestrian who just cut in. What matters in this case is how the algorithms, either in the LiDAR itself or in a separate computing device, could utilise the point cloud 3D reconstruction from the LiDAR data to make proper driving decisions such as braking, accelerating or turning. The evaluation of the performance would be certainly based on that. Eventually, your business model would be built around some unique value propositions enabled by technology and the corresponding solutions. You can then strategically grow your own market advantage with a combination of sustainable user group, partners and investors. If you would like to know more how the NSSN can assist your business in the development of LiDAR or smart sensors, contact us for a coffee chat. We would be glad to assist you further in preparing for the coming sensing era.