OpenECU Deployed on Autonomous Vehicle
Many vehicles now include advanced driver assistance systems (ADAS) such as lane departure monitoring, collision avoidance, backup assistance cameras, blind spot monitoring, and adaptive cruise control. These systems improve safety for current generation vehicles but there is a lot of interest in providing the ultimate safe driving environment with fully autonomous vehicles taking on the full burden of situational awareness.
There are numerous challenges to the advent of the fully autonomous vehicle that must be solved by a large number of stakeholders.
Pi Innovo’s focus, working alongside partners AutonomouStuff and Harbrick, was on architectures for vehicle control systems actuation; one piece in a large and complex puzzle.
Pi Innovo’s OpenECU model based embedded controls platform, sensors for situational awareness from AutonomouStuff, and the PolySync autonomy operating system from Harbrick made for a powerful combination of tools for the development and deployment of ADAS functions and autonomous capabilities.
PolySync is a very flexible set of software tools that provides an API for application developers to create functionality such as lane keeping algorithms, object detection algorithms, and path planning algorithms. This is made possible by abstracting the interface to the numerous possible vehicle sensors, such as radar, lidar, cameras, from the developer, and organizing all the information such that data from any device can be accessed using common function calls, and the user can rely on the data being in a common reference frame with consistent time stamping.
Pi Innovo and Harbrick developed an interface that defines communication flow from the top level autonomy applications, e.g. path planning, to an OpenECU vehicle systems controller that can automatically make the vehicle react. A Kia Soul test vehicle has been modified by Pi Innovo and Harbrick engineers to include brake, steering and throttle actuation by an M220 OpenECU rapid control prototyping controller taking commands over CAN from the main PolySync processing unit carrying out all of the complex calculations required to define requests of those vehicle systems.
Results and Impact
Using OpenECU for embedded controls and PolySync for sensor abstraction and fusion allowed for vehicle autonomy functionality to be prototyped and tested within a day. Taking this approach also provided an accelerated path to production deployment, whether used for small captive fleets that can operate in locations not subject to the regulations and realities of the public highway, or the creation of reference designs for higher volume manufacture.
Model Based Embedded Control
Autonomous Driving Algorithms