How to Make Virtual Prototyping Better Than Designing With Hardware - Parts 1, 2

In automotive, and many other industries, feature sets are now growing beyond one dedicated controller or subsystem and into distributed subsystems. VPs, when combined with models of devices under control (often called plant models), open the door for developers and verification engineers to work with the product in the context of the "full" system.

EE Times Design

Design Process Changes Enabling Rapid Development

A comparison of electronic development in the wireless industry when compared to the automotive industry.

The automotive electronics industry is using the same technologies and methodologies as the wireless industry, but at an introductory pace that lags wireless by about 2 to 3 years ...

Conference Publication, Proceedings Convergence, Society of Automotive Engineers

Home > Solutions > Automotive Systems
Automotive Systems

Automotive Systems

For Automotive OEMs and Tier 1 suppliers

EST’s technologies and methodologies enable the executable specification, architecture, design, verification and validation of distributed, real-time systems having both discrete and continuous behaviours.

In the automotive world, the ESSE Systems Workbench provides a complete model-based specification, design, development, and verification environment that enables software, hardware and plant engineers, individually and in teams, to build entire high fidelity vehicle control systems operating on high fidelity plant models. This includes building functionally and timing accurate real-time control systems, embedded in high fidelity vehicle dynamics models of the target vehicle, that can be driven on accurately modelled roads in accurately modelled traffic.

ESSE Systems Workbench for Automotive Design and Development of Software and Hardware, Verification and Proving
ESSE Systems Workbench for Automotive Design and Development of Software and Hardware, Verification and Proving

Such models can be mathematical – capturing both function and timing – or operational where function and timing of control elements is carried out by software executing on high fidelity ECU models, or some combination of mathematical and operational.

Even when mathematical, the model of control is expected to map to hardware (for instance, an assembly of interacting finite state machines) or software that is able to be executed on some ECU (model and equivalent physical entity). When the nexus between function and timing is broken – for instance, in typical Simulink models – automatic translation from untimed to correctly timed is impossible. It is true that, with considerable manual effort, untimed functional models can be converted to timed functional models and then be used to map to correct hardware and/or software.

Build and Simulate high-fidelity virtual mule vehicles prior to physical prototypes (mules) – save multiple millions of dollars per new and updated vehicle model

The ability to build entire high fidelity vehicle models and simulate them individually or as (autonomous) vehicles operating in traffic enables the testing not just of single, and multiple, network connected ECU + plant subsystems, but the entire vehicle or multiple connected vehicles operating in traffic. This capability provides high fidelity virtual mule vehicles for testing prior to committing to any physical engineering.

Since well-designed models should be highly parameterized, a single high fidelity vehicle model can give rise to many thousands of unique high fidelity vehicle variants. Some parameters will effect structural changes, others functional and timing changes, and others behavioural changes. Now we can simulate thousands of different virtual mule vehicle variants in thousands of traffic scenarios, including scenarios that would have fatal outcomes. It is estimated by vehicle calibration and diagnostic engineers that this capability will enable OEMs to build ~20% less physical mule vehicles for new and updated vehicle model designs. This represents savings in the multiple millions of dollars per vehicle model.

Concurrent real-time systems simulation using multi-core computers

The ESSE Systems Workbench enables simulations to be executed concurrently on separate multi-core computers enabling thousands to millions of investigations to occur in modest amounts of time. This also happens to be the starting point of the empirical process by which the true optimization of vehicle control and plant can be achieved. Optimization processes such as Design of Experiments dramatically reduce the number of simulations required to achieve optimal outcomes.

Entire vehicle models of the calibre discussed above require high fidelity control, plant, network and vehicle dynamics component models. As stated earlier, such models may be mathematical or operational. Interestingly, many vehicle dynamics models are self-contained and present with the capability to drive through sophisticated 2D and 3D roadway infrastructure (for example, Mechanical Simulation Corp’s CarSim models). In actuality, these vehicle models contain high fidelity torque, traction, suspension and vehicle physics models bound with replaceable lower fidelity engine, transmission, braking, steering, body and networking models.

The ESSE Systems Engineering Workbench enables the individual or wholesale substitution of these replaceable models with high fidelity mathematical and/or operational plant and controller models. Such replacement yields high fidelity vehicle models.