CES 2018 Autonomous Driving Tech Review
The term “autonomous driving” arrived in the vernacular at CES 2013, when Audi and Toyota unveiled their self-driving cars. Over the following 12 months major car manufacturers rushed to announce their own autonomous development blueprints or self-driving car prototypes. Google searches for the synonymous term “self-driving” increased sixfold from 2013 to 2017 — with peak search times corresponding with each CES.
At CES 2018, 555 companies or organizations participated in the “Automotive/Vehicle Technology” category, 23% of which were companies using autonomous driving technologies. Featured products can be roughly divided into six categories: Sensors, Perception and Decision, Control, Human-Machine Interaction, V2X, and Miscellaneous.
Sensors are any devices which can provide the vehicle system with raw signals. Generally speaking, this includes GPS, IMU, camera, LiDAR, radar, thermal camera, and so forth. At CES 2018, 23% of the autonomous driving related companies were sensor technology companies, half of which were from Israel. Most of showcase releases were in the LiDAR and thermal camera fields.
- AdaSky showcased its Viper thermal camera which uses heat profiles for obstacle detection. It combines a thermal signal with machine learning to enable the vehicle to detect and understand the environment in all weather conditions.
- RoboSense unveiled its solid state LiDAR with 200 meters range. It also demonstrated its camera and LiDAR signal fusion technology which can detect colours in the real world. Mass production will begin in late 2018 or early 2019.
Perception and Decision
One of the most biggest challenges in core autonomous driving application is perception, which refers to the ability to draw relevant conclusions from the raw sensor data using data processing, data calibration and fusion, scenario reconstruction, localization, and obstacle detection. Because chipsets are the key hardware for data collection and processing in perception, chip manufacturers have also been included in this category.
Decision is the process of making choices based on known data. For example, to help the vehicle take action when data indicates there is an obstacle on the road. At 2018 CES, 33% of autonomous driving related companies were in the perception and decision category, with 20 focused on perception and 22 on decision. Hot topics this year were hardware processor chipsets and Lyft’s Aptiv project.
- Horizon Robotics displayed its new processors Journey and Sunrize. Journey can support the high-performance ADAS system and enable the sensor fusion function to handle multiple camera signals simultaneously. Horizon Robotics cooperated with Intel to provide the L3/L4 autonomous driving solution.
- Aptiv partnered with Lyft to provide an autonomous driving taxi service for 2018 CES attendees. The vehicle navigated busy Las Vegas roads to transport passengers to their destination. Aptiv was responsible for the system’s sensor, perception and decision components.
Control refers to the ability to identify and process decisions into real vehicle actions. For example, if there is a decision ordering the vehicle to brake, control will determine how to activate the brake and steering functions to give passengers the best driving experience. At CES 2018 there were only 12 companies in this category, less 10% of all self-driving related companies. Star attractions at CES 2018 included Xpeng’s G3 SUV 360° camera system, and Nissan’s IMx SUV.
- Xpeng Motors, a new automaker from China, unveiled its G3 SUV. The vehicle is equipped with a 360° camera, 25 sensors, and specific autonomous driving functions to deliver robust autonomous driving capability. G3 will be ready for the public in the spring of 2018.
- Nissan unveiled its IMx, an electrical concept SUV. The car has autonomous driving capability and is integrated with multiple sensors to help the driver monitor blind spots around the vehicle.
HMI is the human-machine interface that connects an operator to the controller in a computer system. This comprises a variety of services and functions to support autonomous driving, for example, voice assistants, smart alarms, and in-car interfaces. There were 16 companies working in this area at CES 2018. Among the most interesting showcases was Autolabs’ virtual assistant project and Steering AI’s smart alarm system.
- German Autolabs brought its first virtual assistant to CES 2018. It has gesture and voice control and can actively assist the driver. The virtual assistant can be easily accessed by third party developers and integrated with additional functions in the future.
- Steering AI from China displayed its smart steering cover which can transmit an alarm if it detects driver is falling asleep. The devices can be quickly installed in fleets and collect driver information for fleet management.
V2X (Vehicle-to-Everything) indicates the communication system between a vehicle and anything else. The category includes V2I (Vehicle-to-Infrastructure), V2V (Vehicle-to-Vehicle), V2P (Vehicle-to-Pedestrian), V2D (Vehicle-to-Device) and V2G (Vehicle-to-Grid). Fifteen V2X companies released blueprints or new products at CES 2018. Especially noteworthy were Qualcomm’s V2X network and RoadEyes’ resSMART camera.
- Qualcomm‘s Cellular Vehicle-to-Everything technology C-V2X is a unified connectivity platform that enables and enhances the vehicle’s LTE connection. C-V2X is designed for high-speed communication and will be ready for commercialization in 2019.
- RoadEyes showed its resSMART camera, which can integrate with smartphones and record trip information such as surroundings, and road and traffic conditions. The built-in GPS will tag the footprint automatically. This device also has a function enabling drivers to share their journey data with other users.
Various autonomous driving technologies cannot be pigeonholed into the above five categories. Baidu’s self-driving development platform Apollo is one of them. Apollo 2.0 is a major release that not only changes the way autonomous driving technologies are developed but also forms a worldwide ecosystem for this technology.
- Baidu released Apollo 2.0, an updated open-source autonomous driving platform. Apollo is a framework to assist in the development of autonomous driving technologies, for example it provides a test module so developers won’t need to build one from scratch.
- Regulus Cyber is the first company to focus on sensor and communication security in the connected world. It aims to protect vehicles, drones and robotics against malicious attacks. Their GPS SP system can protect GPS devices from hijacking.
Geographic Info on CES Self-driving companies
American companies remained major players in the self-driving field at CES 2018, accounting for just under 30% of the total. Most were from California, suggesting the next autonomous driving wave will most likely emerge from Silicon Valley. Meanwhile, over 40% of sensor companies were Israeli.
CES 2018 data also indicates strong growth for self-driving tech companies in Asia, as more Japanese, South Korea and Chinese companies are entering the market. In China dozens of startups and high-tech companies have shown their capabilities in this area.
The Future of Autonomous Driving
Based on trends at CES 2018, it’s easy to see that competition is heating up in the autonomous driving industry, and more investment will be poured into this area. However, such investment may focus on specific breakthrough technology rather than the general self-driving field. LiDAR and Autonomous Driving Processor Chipsets could be two major battlegrounds. Moreover, based on the various partnerships announced at CES, we can envision self-driving competition evolving to the ecosystem level. Dozens of companies aim to have their full set self-driving mass production vehicles on the road by the end of 2018 or early 2019.
Although autonomous driving technology has been under development for some time now, 2018 may be the first year that self-driving vehicles actually appear in our everyday lives.
Contributing Analyst: Alex Chen | Editor: Michael Sarazen