Autoware – open source self-driving coming of age?
Autonomous cars have been one of the most interesting technical challenges in the last few years. The problem seems to have attracted the smartest brains across technical universities and industries alike. In fact, there was even a time when self-driving car engineers’ salaries had become a discussion point as they had shot up so much that only a few companies with deep pockets or highly funded startups could dare to work on the self-driving problem.
But, things seem to have changed in the last couple of years. Open source solutions have caught up and now it is much easier to build a self-driving car prototype even for startups or University projects. One latest and important step in this direction was the recent funding of $100 million to Tier IV, the open-source self-driving startup based in Japan to facilitate commercialization of self-driving technology in private areas, depopulated areas and urban areas. The funding round was led by Sompo Japan Nipponkoa Insurcane Inc., with participation from existing investors – Yamaha Motors, KDDI Corporation, JAFCO, and Aisan Technology.
Tier IV was founded in December 2015 and develops and maintains Autoware Foundation along with other premier members. Autoware was officially launched in December 2018 and has been accepting new members since. It has been supported by companies like LG, Toyota’s TRI-AD (Research focus on autonomous driving), ARM, Apex.ai and many other startups. It started with autoware.ai which is built on ROS 1 and claims that it has been adopted by over 200 companies in their projects.
Autoware now works on autoware.io – interface software like device drivers for sensors, controllers, etc. and autoware.auto which is a complete rewriting of autoware.ai on ROS 2 with an early focus on automated valet parking and autonomous depot maneuvering but has much larger plans in the future. The focus, however, remains on autonomous driving in certain constrained areas or speed constraints and target mobility-as-a-service market rather than level 5 autonomous cars. This seems to be the realization for many car companies as this is the easiest way to commercialization in the early stages of the technology.
Autoware is not the first open-source project in autonomous driving. Baidu announced the Apollo Open Platform in April 2017 and the project has attracted many marque partners since including most Chinese automotive OEMs, Honda, JLR, Daimler, ZF, Continental, WABCO and many others. It released Apollo 1.0 in July, 2017 with closed environment capabilities and has just released the Apollo 5.0 on 2nd July, 2019 for volume production, geo-fenced autonomous driving. It plans to release the full autonomy version by 2021.
Other important solutions include Renovo Auto’s AWare which is a more commercial version based on open-source software as well as Comma.ai’s solution. This is not just limited to the autonomous software stack. Open Motors (previously OSVehicle) develops an open EV hardware platform for autonomous driving targeting the MaaS market. The company is backed by Groupe Renault.
A similar approach is followed by UK based startup StreetDrone, which provide organisations the AV platform to develop their own self-driving applications, without having to invest the time and money creating self-driving vehicles. They also assist in the deployment of open source software stack Autoware, as a founding member of the Autoware Foundation.
It is also very clear here that these open source projects have a distinct major company backer, which is a clear sign that these companies are eyeing the value creation for themselves. This makes it difficult for companies to adopt these solutions directly. But, these solutions seem to make prototyping an initial solution very easy. This is definitely great democratization of self-driving technology as new startups and other companies can create a working prototype of solutions very fast. This also helps in expanding the adjacent technology areas and service providers as they can easily test their solutions. Also, this industry-wide collaboration in sharing the technology will enable to lift the safety standardization for autonomous technology as a whole.
However, some people might find the idea of an open-source self-driving car disquieting, fraught with potential for security vulnerabilities and misuse. For example, Waymo claims that developing its technology internally allows for more efficiency and tighter integration. All together open source developments are driving towards collaborative R&D efforts in developing autonomous technology and creating a development bed for upcoming startups.
To deep dive and stay continuously updated about the most recent global innovations in Autonomous Vehicles and learn more about applications in your industry, test drive WhatNext now!