A Rapidly Changing World
We have been engulfed with ideas, spirited discussions and emphatic opinions on what it would take to create the city of tomorrow. A big facet of this vision is self-driving cars, a technology that will impact many aspects of our day-to-day lives – the way we travel, our lifestyle, the creation of new industries and employment, while disrupting others. The world is changing very rapidly, and we need to innovate to adapt to this new, emerging situation. Vehicle ownership in large cities is on the decline as inhabitants find it easier, more convenient and cost-effective to use ride hailing services. Shopping malls and departmental stores are fighting for survival with the increase in online shopping and hectic lifestyles are increasing our reliance on food delivery services. With more online sales, innovative means of package distribution and delivery is becoming major technology trend.
With all of this in mind, it’s time to look beyond the potential and do a status check on where we really stand with the implementation of truly autonomous vehicles (AVs) and what would it take to move autonomous tech from ‘data collection mode’ to mass deployment for everyday use.
Building Trust in Autonomous Vehicles
Let’s start by looking at how far we have come in the evolution of self-driving technology from the days of the first Waymo car. Looking at the US market where the state of California (the only state where they are enforced by The Department of Motor Vehicles (DMV) to release reports if testing AVs), twenty-eight companies reported results. In total these companies drove greater than two million miles in autonomous mode and encountered 73,550 disengagements (i.e. the moment, the system hands back control to a safety driver, or the safety driver must interfere for safe operation of the automobile). The average disengagements per mile were at five per 1000 miles (or one disengagement every 200 miles). With this level of consistency, California DMV relaxed its rule on having a safety driver when the car is driving in autonomous mode.
This represents a significant shift, and so far, the only company that has been awarded this license is Waymo. California recently also opened the door for testing of light duty autonomous trucks, a category that includes minivans, pickup trucks, utility vans, and step vans. This visibility into the reporting, while providing a huge advantage in educating the citizens, also necessitates the need for more details to draw meaningful conclusions. For example, if a company reports on a disengagement happening due to perception issues, the current DMV rules do not mandate the company to report on which object failed (or is repeatedly failing), identification or classification. In the absence of a standard process or methodology to verify the accuracy of data, the efficacy of the reports is questionable in terms of increasing the public’s trust in the technology.
Across the Pacific, China is slated to become the largest market for AVs. Per a McKinsey report: “autonomous vehicles could account for as much as 66 percent of the passenger-kilometers traveled in 2040, generating market revenues of $1.1 trillion from mobility services and $0.9 trillion from sales of autonomous vehicles by that year.”
In December 2017, Beijing became the first city to release policies for self-driving vehicles on the city's roads, this was followed by Shanghai. Subsequently, in February 2018, Beijing opened the first testing facility for autonomous vehicles. The innovative 133,000 square meter facility mimics real life testing environments with scores of static and dynamic traffic scenarios experienced on urban and rural roads. The Chinese government has invested heavily and committed to large sums of money and resources in encouraging companies to develop the most advanced technologies to grow in this market space and at a fast pace. China wants to position itself as a technology leader, spearheading the AV market.
The European Union (EU) also clearly acknowledges the benefits AVs hold and how pervasive their impact can be. There is robust cooperation across member states ensuring wide deployment with consistency and commonality of self-driving technology across the various nations. This is apparent in The Vienna Convention on Road Traffic that necessitates consistent traffic regulation across borders. It further reinforces a consistency in the creation of highways connecting cities making it conducive for AV testing. Furthermore, the EU rolled out a plan that will see driverless trucks operating between cities in as little as two years. Reduced transit times and fuel costs are a few of the many expected benefits.
Let’s take a more granular look at some of the countries within the EU. Germany, often considered the mecca of traditional automotive companies, was one of the first countries to legalize AVs on roads - albeit with safety drivers. The Netherlands is by far, leading the pack, not just in Europe, but globally. The country now allows testing of self-driving vehicles without the presence and control of human drivers. Sweden has provided the platform for test programs involving mini buses with multiple companies undertaking extensive and ambitious roll-outs of Autonomous buses. Owning a non-autonomous car will soon be like owning a horse,” said Carlos Moedas, the EU commissioner for research, science and innovation. With all these countries working under a common framework, Europe is creating a safer and more efficient road network through the development of a nationwide strategy to integrate AVs.
Sensors Are The Eyes of Autonomous Vehicles
Sensor technology is a critical component enabling AVs. They provide the ability to detect objects in the path of the vehicle and act as the eyes, providing data upon which automated decisions can be made. Let’s look at a few sensors at the center of AV technology:
Cameras are the master of classification and texture interpretation, and today are the most commercially viable sensor option. CMOS image sensors - a major component of the camera - have come a long way from the days of pretty pictures and selfies.
These sensors have evolved from ‘viewing only’ sensors, to viewing plus sensing. They are the primary sensor used for functions such as front view Advanced Driver Assistance (ADAS) and traffic light recognition. The sensor data collected relays a vital feed into the decision-making process of the car. CMOS image sensors used in AVs today are equipped with very advanced functionality enabling them to efficiently handle all driving scenarios in all types of environmental conditions.
CMOS images sensors act as eyes of the Autonomous Vehicle
High dynamic range (HDR) is a key and critical specification parameter of camera sensors. This is the ability of the sensor to see between the darkest part and brightest part of the scene in the same frame without over or under saturation. In other words, the sensor needs to be impervious to changing ambient light conditions, such as suddenly been exposed to a bright sunny angle or reflection, or moving in and out of a tunnel in daylight. Companies like ON Semiconductor provide imagers with the highest dynamic range (HDR) for safe dependable use on AV applications.
Another trend for autonomous driving is higher resolution sensors. This is being driven by the need for autonomous cars to see further and at the same time can detect objects of all sizes. Essentially, this gives the ability to increase the number of pixels per degree on the object making the job of self-driving algorithms more efficient with both object detection and classification.
For level three AV cars and beyond there is a heavy dependence on hardware and software to make decisions that would otherwise be taken by humans; this can also make vehicles more susceptible to cyber-attack or hacking. This has prompted work on the development of numerous protection and security features including cybersecurity, being added at the sensor level. In recognition of this challenge and threat, ON Semiconductor is introducing the first automotive grade imager with fully integrated cybersecurity features.
Companies working in the field of LiDAR have attracted a lot of attention from investors as well. This technology is highly desirable for self-driving cars, though can come with a very high price tag. The ability of LiDAR to create a 3D point cloud sets it apart from other sensing technologies. Waymo changed the playing field with the introduction of its in-house built LiDAR that not only detects pedestrians but also can assess the direction pedestrians are facing and what direction they will move towards. This also gives the cars the ability to see hand signals from pedestrians or cyclists, and then adjust driving behavior accordingly. Range, resolution and rotation/frame rate are other critical parameters, when selecting LiDAR. The latest LiDAR technology can see more than 200 meters. The idea being not only to detect an object on the road, but also can classify, and provide feedback to the decision-making system and change the course of the vehicle as needed. Reflectivity of objects plays a vital role too. While current LiDAR can detect objects with high reflectivity at distance, work still needs to be done on detecting extremely low reflectivity at long distances.
Radar is another sensor type that is frequently discussed and is critical for the future of autonomous driving. The unique benefits of radar are apparent when operating in the dark and in all weather conditions. Millimeter wave (mmWave) is a special class of radar technology that uses short wavelength electromagnetic waves and is starting to become popular for autonomous driving applications.
By capturing the reflected signal, a radar system can determine the range, velocity and angle of objects. A mmWave system operating at between 76 and 81 GHz (with a corresponding wavelength of about 4mm), will have the ability to detect movements that are a fraction of a millimeter. This provides an accuracy level never seen before in the automotive space. It allows precise measurements between the vehicle and surrounding vehicles and objects. Another improvement is seen in the ability to make the design more compact and integrated, without the need for large, visible antennas.
AVs are one of the most highly anticipated innovations with a potential impact that will change the lives of millions across the globe. As this technology becomes accessible to the public, the overall effects – whether positive or negative – are sure to be significant. One thing is certain – we are in for an interesting ride