MotionWise adds data flow driven scheduling

MotionWise adds data flow driven scheduling

Technology News |
By Nick Flaherty

The MotionWise Schedule tool handles bounded end-to-end latencies to schedule software tasks and communication across the network where safety and real-time behaviour matter

TTTech Auto has added support for dataflow driven scheduling to the MotionWise scheduler to handle dataflow dependencies and allows to dynamically switch between different dataflows. Modeling is simple and efficiently based on Directed Acyclic Graphs (DAG) with parallel execution on multiple CPU cores.

Computation chains can be easily modeled and end-to-end latency bounds (from sense to act) are enforced across multiple CPU cores in system on chip devices. It supports AUTOSAR, POSIX, Data Distribution Service (DDS) and Time Sensitive Networking (TSN).

As the industry shifts towards software-defined vehicles (SDVs), characterized by an increasing number of software components, intricate dependencies, and more complex underlying hardware resources, a new layer of complexity has emerged. TTTech Auto is part of the NXP ecosystem for SDV software.

Ensuring consistent end-to-end timing from sensors to actuators is more crucial and challenging than ever but exhaustive testing of all possible states has become unfeasible as traditional correct-by-testing methodologies are ineffective.

The task of efficiently allocating software components to resources, while adhering to design constraints, has become increasingly complex so maintaining system integrity under scenarios of CPU and network overloads requires new tools.

The MotionWise Schedule tool uses a correct-by-design approach with TTech Auto’s scheduling algorithms to  ensure that software components and communication are  mapped to the hardware resources so that safety and real-time behaviour can be achieved, says Dr. Stefan Poledna, CTO and Co-Founder of TTTech Auto.

This builds on the TTTech Auto MotionWise safety middleware designed for distributed architectures requiring safety and real-time properties. Today, MotionWise is deployed in more than 2 million vehicles, with a secured pipeline of more than 9 million in the years to come.

In SDV projects, manual configuration of schedulers often leads to a cycle of repeating testing and iterations. With each code alteration or addition of a new feature, with every new software release the process re-starts, causing delays in start of production (SOP) timelines and escalating costs. MotionWise Schedule can circumvent this through the scheduling concept encompassing workload separation based on criticality, task dependency scheduling, and failure handling.

An embedded stack implements these functionalities within the vehicle.

 “This method significantly reduces the time required for software releases, encompassing design, testing and validation phases. The “MotionWise Schedule” runtime enforces configuration outcomes, offering a safe execution environment that includes safe real-time and run-time scheduling, guaranteed end-to-end timing, Freedom From Interference (FFI) for mixed-criticality systems, and deterministic execution among other properties,“ he said.

MotionWise Schedule is the first modular solution from the new MotionWise family. It contains all the technology developed for MotionWise in a modular and flexible product, enabling a seamless integration with standardized based architectures as well as custom ones.

It allows scheduling of tasks and network communication to achieve higher resource utilization and performance providing a correct-by-design approach. It seamlessly maps tasks to CPU cores and defines the required network configuration for Time Sensitive Networking (TSN). Thus, it enables highly optimized multi-CPU core and multi-SoC solutions.

Time-Triggered execution provides a time-boxed isolation between tasks, enabling mix-criticality and composability. This results in lower testing and verification efforts for real-time tasks, determinism and execution repeatability, as well as faster system integration.

Task monitoring is provided to gain comprehensive insights into task performance and sequencing allowing for real-time analysis and system level optimization.

Time synchronization across multiple SoCs and Ethernet TSN is provided for time-triggered execution and communication, minimized latency, and real-time performance analysis.

Intuitive tooling allows to visualize schedules and supports comparisons of plans versus actual execution. This enables faster and more efficient pinpointing of resource issues and supports faster system level optimizations.

MotionWise Schedule abstracts software functionality from the hardware with its powerful tooling and run-time. It supports mixed-criticality systems and supports incremental updates with vastly reduced re-validation efforts.

If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :    eeNews on Google News


Linked Articles