Introduction to Autoware.
Building a self-driving stack from scratch is complex, time consuming and extremely resource intensive. Building a team to take on such an ambitious task requires scarce technical skills, large amounts of know-how, and a deep understanding of the various market factors, including regulations and safety standards.
To get your self-driving project up and running at speed, StreetDrone can provide bespoke implementations of an open source self-driving stack to accelerate this process and push your project to the forefront of the new mobility sector.
Benefits of Open Source Self-Driving
Autoware provides our customers with the openness, flexibility and future development scope they require to explore the self-driving space, and get their autonomous vehicle development programme kickstarted at an accelerated rate:
Open source self-driving software provides an accelerated starting point, saving you from re-inventing the AI-controlled wheel. Rather than spending all of your time building up general functionality, you can focus on the area which interests you most, and where you intend to build core IP and business models.
Open source provides flexibility. Very permissive licencing means that the stack can be re-used, rebadged and re-marketed for commercial purposes. It can also be forked to a designated, offline version to act as a stable reference stack should the user wish.
The power of the self-driving community is working hard to ensure the software is constantly evolving and improving. Expert organisations and individuals from all over the world are working on improving the stack’s reliability and functionality, to the collective mutual benefit of all involved collaborators.
The nature of open source software means that all data is open data. Involvement in the open source community removes reliance on “black box solutions” that provide functionality without insight. Delving into the open source world allows you to extract data which is of specific relevance and value to your application, usecase and business model.
Active involvement in the open source software community rapidly expands your partner network. Should you require specific expertise, from academia and tech start-ups to sensor manufacturers and OEMs, the community allows you to make valuable, targeted connections. This includes implementation partners such as StreetDrone.
Open source allows rapid scalability. Once the software stack has been adapted to your specific usecase, expect the potential to grow fleet size and number of deployments rapidly, without reliance on costly 3rd party licencing.
Continual evolution ensures you remain planted at the leading edge of the self-driving industry, with a community discussing and building complementary, stack- optimised tools and hardware.
What is Autoware?
Autoware is a Robot Operating System (ROS) based open source software distribution developed by a worldwide community of contributors, built by Nagoya university, and now maintained by a global group of collaborators, the Autoware Foundation.
The software distribution is broken into 3 main, continually evolving projects:
Autoware.AI - based on ROS, Autoware.AI is the original form of Autoware, currently used primarily for academic/ early stage R&D purposes, and as such has not yet been widely tested on public roads.
Key Features include:
Localisation - 3D Mapping, SLAM algorithms, GNSS and IMU Support
Detection - Fusion of Camera, LiDAR sensors
Prediction and Planning - deep neural networks, rule based systems
Autoware.IO - a project aimed at improving the interfacing of various 3rd software / hardware party packages including proprietary systems. Examples include device drivers for sensors, by-wire controllers for vehicles, and hardware-dependent programs for SoC boards
Autoware.Auto - A project aiming to move the functionality of Autoware.AI for the automotive world. Autoware.Auto migrates the key features of Autoware.AI from ROS to ROS2 to make the system truly realtime, and capable of running on safety critical hardware.
StreetDrone and Autoware
StreetDrone joined as a premium founding member of the Autoware Foundation in late 2018 before its launch early in 2019. StreetDrone’s role is to provide autonomous-ready vehicle platforms, complementary tools (open source ROS vehicle nodes, simulator models, data services) as well as implementation support.
StreetDrone’s vehicles have full L2 Drive-By-Wire safety case, with associated functional safety alongside integrated sensors and computation hardware, built into form factors for a range of usecases (including R&D, robotaxi, and first/last mile mobility/logistics) To assist in the ease of implementation,
StreetDrone is building its “vehicle interfaces” to be out of the box compatible with Autoware (read more about our ROS node - Vehicle Interface )
StreetDrone customers are already making headway with Autoware combined with StreetDrone vehicles:
Parkopedia / Autonomous Valet Parking – parking specialist Parkopedia are deploying Autoware on their StreetDrone Twizy as part of an InnovateUK funded consortium including the Transport Systems Catapult, Surrey University and Admiral Insurance Group. The project addresses the problem of automated parking, including GNSS-denied locations (such as multistory). Read more.
HKPC – Hong Kong Productivity Council are deploying Autoware on a StreetDrone Twizy as a reference vehicle architecture in their role as innovation specialists in Hong Kong. Read more
Millbrook Proving Ground have three Autoware compatible vehicles as part of their role in supplying CAV testing services for Testbed UK. Read more