As a matter of fact, autonomous sedan car is not a miracle now, where tech companies are hugely investing and refining the Lidar technology. Certainly, the sensor behind the autonomous vehicle development.
LiDAR is an acronym for light detection and ranging. In other words, it uses light-sensing technology in the form of a pulsed laser beam which is directed towards the object to measure the distance.
The technology was developed in the 1960s and mostly used in airplanes in the past. However, it started gaining attention after the United States of Defense invested in developing a GPS technology. US defense launched a first GPS prototype by 1978.
Lidar was invented in 1961 by Hughes Aircraft Company. They also made a laser in 1960. After that, the technology was utilized by the United States for Apollo-15 mission in 1971 to scan lunar surface.
Let’s see different aspects of Lidar. We will go through the below important points to understand its working and application in the auto industry in brief.
- Lidar and its parallel sensors (Sonar and Radar)
- How Lidar Works?
- Lidar Components (GPS, IMU, and Laser)
- Secondary Components in-short (Gyroscope, Accelerometer, and Photo-diodes)
- Lidar Types based on Functionality
- What is Lidar Intensity?
- Investment and Application in Automobile
Lidar and Its parallel sensors
Earlier, two other technologies were ruling the world due to there affordability and precision. I.e Sonar and Radar. All these sensors are working on the principle of echolocation (Use of echoes to measure the distance of the objects).
Sonar is a sound navigation and ranging technology, However, animals like Dolphins and bats are using sonar to navigate and measure distances from the ages.
Let’s see why sonar is good at underwater?. Firstly, sound waves uses its vibrational nature to find out the distance between objects. Secondly, they travel faster than light and radio waves inside water.
Radar is a radio direction and ranging technology. However, radio waves are used to measure the distance between the targets. Radar is also used to detect objects in the rough weather conditions with the accuracy of 2cm.
Without delay, it emits short radio pulses in the direction of the antenna. Radar was developed during World War II.
How does Lidar work?
Lidar is a dynamic remote sensing methodology. Since, the pulsed beam fired (150,000 pulses per second) from the sensor, collide with the desired objects and reflect back towards the sensor receiver. This process is also called as a time of flight.
You can see the simple block diagram on how does lidar work and its different working components.
As a result, the reflected light is collected by the receiver sensor located on the Lidar module.
In reality, the Lidar sensor and detector are continuously rotating to capture surrounding data. While the Lidar sensor is installed on the top and side of the vehicle. The above 2D Lidar image is a conceptual schematic of Lidar Data, since sensor capture and displays the received data in different colors (wavelength).
Meanwhile, we can use the basic velocity measurement equation to find out the distance traveled by the laser pulse. Velocity = Distance / Time (Velocity = Light travel velocity [186,000 miles per second], Time = Pulse travel time)
Distance = (Speed of Light x Time of Flight) / 2
A pulsed beam is a non-continuous form of optical energy hence, pulses are appearing repetitively in a certain time of period. In fact, pulse energy calculates in Joules.
In short, the Lidar technology helps to map out any surrounding environment. After all, it may be farming, buildings, water, forests, etc.
It consists of 3 distinct components especially, that help to complete the function of data collection.
- Laser device- It significantly emits the light-pulses towards the targeted object and helps to find out the delay time between the light-pulse transmission and its reception back.
- Global Position System (GPS) – Notably, GPS helps to identify the Lidar sensor location or instrument.
- Inertial Measurement Unit- It consists of a gyroscope, accelerometer, and magnetometer which also help to detect linear and angular elevations. Moreover, It is a form of a 3D scanner.
Secondary Lidar components
Gyroscope is used to measure the angular velocity while maintaining the stability of the objects. The basic working principle behind the gyroscope is that it has a spinning wheel that rotates around the central axis and resists to change in orientation regardless of its position in the space. However, it maintains the identical position before the spinning disc drops the speed.
Whereas, the accelerometer is the electro-mechanical device used to measure the acceleration. Acceleration is nothing but the speed (Velocity) divided by time. The accelerometer is widely used to take preventive actions like airbag deployments during the car crash.
The best example is when you rotate your smartphone from portrait mode to landscape mode. However, the physical action of phone tilting is sensed by the accelerometer and initiates the rotation process of the screen.
The magnetometer is a magnetic sensor which is used to measure magnetic field intensity and particularly used in aircraft altitude determination systems. They also use the earth’s magnetic field as a reference.
Photo-diodes are photosensors that can absorb light and convert it into a current. While they consist of optical filters and built-in lenses. A solar panel is also a big photo-diode.
Photodetectors (photo-diodes) and receivers which help to collect the different wavelength of lights convert it into a different category as well as display reports in different color formats.
The data collected by Lidar in the form of discrete data points like X, Y and Z coordinates converted into 3D visuals which also helps to explain the characteristics, shapes, and density of the objects.
However, two to three data points might be collected by an average single pulse. As more and more data points collected, you will obviously get denser 3D models of the object or the surrounding environment.
What are the Lidar types?
There are 2 types of Lidar based on their functionality.
Airborne Lidar (500-2000m)– They always positioned in the air like airplanes, helicopters, drones, etc. All in all, it will help to collect the larger data points and capture more coverage of the area.
Overall, the Lidar range may vary from 50m to 250m or more.
Airborne Lidar divided into 2 types.
- Topographic Lidar uses a near-infrared laser to map the land or collect the data.
- Bathymetric Lidar is used to carry out the mapping inside of the water. Bathymetric is a terminology used to represent the ocean’s depth relative to sea level and mostly used to map out underwater terrain.
Terrestrial Lidar – it generally mounted on earth’s surface or moving objects like cars to collect the nearby data. Terrestrial Lidar is an important element of autonomous vehicles while it is replacing human drivers altogether with the help of artificial intelligence.
They are further subdivided into 2 types-
- Mobile Terrestrial
- Static Terrestrial
After all, Lidar technology helps scientists, young automotive startups, and space agencies to measure and examine man-made and natural object’s accuracy, precision, and flexibility.
“.las” is the commonly used file format to store Lidar data until now. Presently, the American Society of Photogrammetry and Remote sensing (ASPRS) supports this format.
The light pulse intensity is eventually helpful to construct the conceptual 3D models of the environment quickly. The sensor module collects the reflected pulses back from the object and records intensity at that point straightaway.
Intensity is further dependent upon the surface texture of the object. Additionally, the receiver, laser beam, incident angle, and range of the Lidar, etc. affect the intensity of the Lidar.
If the angle of reflection increases, therefore, returning energy of the laser pulse decreases.
In the past Lidar costs up to $75000. In 2019 the average Lidar unit cost was around $30,000 per module. Certainly, Lidar is an expensive component to assemble into automotive vehicles. Occasionally, the wide use of this technology might reduce the production cost of the sensor up to $500 by 2030.
Tesla CEO Elon Musk has a different view on Lidar technology and said it is a “fool’s errand,” anyone relying on Lidar is doomed. Doomed. Expensive sensors that are unnecessary. It’s like having a whole bunch of expensive appendices.” Musk argued that a camera and radar-based system with powerful strong AI will solve the problem.
Investment in Automobile
First, It is certainly a key component of any autonomous or self-driving vehicle. Today auto OEM companies including Toyota, Audi, FCA, Honda, and Tesla along with ride-sharing startups like Uber, Lyft, Grab, Didi Chuxing, and Google’s Waymo significantly investing billions of dollars to develop fully autonomous (Level 5) passenger vehicles.
GM Cruise has developed a fully autonomous electric prototype however, they aim to achieve level-3 autonomy by 2021.
Honda motor indeed invested around $2.75 billion in GM Cruise over the coming decade. Therefore, Honda will acquire approximately 5.7% equity in GM Cruise unit for $750 million.
Softbank also collaborated with Toyota to form a new entity called Monet Technologies. So they can develop autonomous mobility to start their own ride-sharing services with no human drivers.
Meanwhile, Toyota invested around $500 million into the Uber to develop the self-driving tech. By the time, the platform will be utilized on Toyota vehicles to power the Uber ride-sharing network around the world in the future.
Application in the autonomous vehicle
We are driving the cars with no automation today also termed as Level Zero. With Level One, Vehicle will assist the Driver in some cases like putting extra pressure to stop the vehicles in emergency. Level 2 & 3 automation give more control to vehicle with driver assistance in accelerating, braking and monitoring the surrounding situations.
Level 4&5 vehicles able to steer, break and accelerate itself and no human assistance required.
Gm Cruise has developed a fully autonomous electric prototype in last 2 years and aims to achieve level-3 autonomy by 2021.
Honda motors will invest around $2.75 billion in GM Cruise over the next 12 years of span. $750 million will be invested upfront to acquire 5.7% equity in GM Cruise unit.
Softbank also tied up with Toyota to form new entity Monet Technologies to develop autonomous mobility development like ride sharing giant Uber but without drivers.
Toyota invested around $500 million into Uber to develop the self driving tech and platform will be utilized on Toyota vehicles to power to Uber ride sharing network around the world.
The image shown at the start of this post shows LiDAR with other important components required to make any vehicle autonomous. Few companies developing the separate retrofit kits which can externally added to any current available vehicles to make it autonomous.
Other companies are separately developing a complete robot as a passenger car like Zoox.