Hokuyo Automation Boosts 4D LiDAR Leveraging SiLC’s FMCW Single Chip Technology
Hokuyo Automation and SiLC Technologies are teaming up to push 4D LiDAR to mainstream using frequency modulated continuous wave technology (FMCW). However, what is 4D LiDAR and SiLC's chip solution?
As the push for autonomous vehicles remains at the forefront of many automotive manufacturers' minds, LiDAR innovations keep popping up.
One recent advancement comes from Hokuyo Automation, which is a leading manufacturer of sensor and automation technology. Joining up with SiLC Technologies (SiLC), a machine vision developer on a new project to establish 4D light detection and ranging (LiDAR) solutions, Hyokuyo hopes to improve machine vision systems for targeted industrial automation and robotics applications.
In this article, let's dive into the journey to create machines with human-like vision, SiLC's latest chip solution, and finally, the differences between 4D and 3D LiDAR.
The Quest for Machines to Perceive as Humans
As humans, the ability to perceive and express different emotions in various ways comes naturally; however, trying to create the same responses within machines is a hurdle many researchers and companies are attempting to tackle, especially when it comes to industrial and automotive areas.
In industrial robotics, automotive sensory, and LiDAR systems, machine vision (MV) technology is commonly used. MV allows for specific cameras to provide automatic analysis of objects at short and long distances. SiLC, a leading developer of MV solutions, has developed technology that encompasses "Smart Vision," the next advancement in machine vision.
SiLC's 4D LiDAR Smart Vision chip. Image used courtesy of SiLC
As mentioned previously, SiLC and Hokuyo Automation are teaming up to bring SiLC's 4D+ smart vision chip into mass production. Over the years, Hokuyo has established a slew of LiDAR and obstacle detection devices to address the many challenges in improving autonomous vehicle sensory. With the clientele and the extensive LiDAR and obstacle detection portfolio of Hokuyo, SiLC could reach mass production at an accelerated rate.
Though both companies aim to leverage the many benefits each company brings to the table, one question that remains is what SiLC's technology is?
4D LiDAR Machine Vision Chip
SiLC has developed the industry's first fully integrated coherent 4D LiDAR chip based on laser technology using FMCW.
An example of FMCW in a LiDAR system. Image used courtesy of Zhang and UC Berkeley
Though you might have heard the buzz revolving around coherent 3D LiDAR, you may not have heard about 4D LiDAR. In general, 4D LiDAR builds off the FMCW technology based on 3D vision but adds a critical vector in terms of measurements to move away for a 3D module into the 4th dimension.
Through 4D, a device could involve polarization intensity, velocity, distance/range, and resolution to allow for machine sensory to get closer to matching human perception.
The technology behind the 4D+ vision chip integrates all required LiDAR functionality, such as a coherent light source and optical signal processing. However, SiLC's solution also enables additional information extracted from the returning photons before converting to electrons.
The transmitted and received optical wavelengths travel through coherent mixing and amplification detector that blocks LiDAR interference such as light refractions and sunlight. This photodetector then conducts the Fast Fourier transform to have the range and velocity signals extracted from the received waveform.
Now that a bit more light has been shined on 4D LiDAR, how does it compare to 3D LiDAR?
Comparing 4D vs 3D LiDAR
Current 3D vision-based LiDAR systems are designed with Time of Flight (ToF) sensors. These sensors operate at around visible light wavelengths of approximately 905 nm and are considered safe regarding human interaction.
Example light wavelengths. Screenshot used courtesy of SiLC
ToF's functionality consists of sending laser-pulses every micro-second to directly measure the time delay between the pulse sent out towards objects and as it returns.
One drawback of 3D ToF solutions is solar interference. Daylight is a challenge when performing ToF detection since light can reflect or refract on certain moving objects, making it hard to determine the distance from the sensor to the object. Another limitation to ToF is range. ToF can't measure at the same level of accuracy once an object is close to 1 km away.
By introducing 4D solutions, the wavelengths push 1550 nm, improving eye safety and mitigating less solar interference. Calculating the range or distance of an object is a direct function of the frequency shift of the returning signal, while velocity adds the frequency shift of the returning signal. However, if they travel simultaneously, called a 'dual chirp,' this resolves range and velocity. Additionally, the FMCW operates lower than a 1 kW of power than pulsed 3D LiDAR, operating at 1550 nm wavelength, which increases eye safety and avoids the challenges of operating in daylight.
However, ToF measurement is still sought after in the automotive industry since it has been well-established for decades. ToF cameras work well with advanced driver-assistance systems (ADAS) sensors that help drivers with lane detection, blindspot, and close proximity object detection.
At the end of the day, both 3D and 4D have weaknesses and strengths. The costs for mass production of FMCW-based LiDAR ICs may limit the supply curve since it is still in the beginning phases of development. What could help FMCW breakout as the go-to measurement principle is being on a single chip and, in the long run, for 4D to propel past 3D ToF detection solutions.