Mehran Mestchian, MathWorks
Engine downsizing using turbochargers is a key strategy for lowering emissions and increasing fuel economy. However, engineers face great challenges in designing these control systems due to their nonlinearities, multivariable nature, system coupling, and calibration complexity. This presentation describes a model-based solution to this complicated control problem using advanced concepts such as inverse model control and quantitative feedback design with a calibration toolset built on System Identification Toolbox and Control System Toolbox™. Using this approach for designing a robust high-performance control system represents a paradigm shift from test cell calibration to a virtual desktop environment.
Yue-Yun Wang, General Motors Company
The current production version of the Bosch OCS would not exist if it had not been for Model-Based Design with MATLAB and Simulink products. This presentation is an account of how this approach contributed to the successful evolution of the Bosch OCS from concept to production. In particular, Model-Based Design was instrumental in providing a solution that was significantly more efficient than conventional heavyweight, software development methods while also enhancing quality.
Russell Watts, Robert Bosch LLC
Generating production embedded software requires a well-defined environment for Model-Based Design. This environment ensures the model complies with industry and company standards. It enforces standardized model architectures, predefines interfaces, and allocates all artifacts in accordance with project partitioning and software architecture. The environment also insulates model developers from software specifics that are needed for creation of production code. This presentation identifies the components of the Simulink based Model-Based Design environment that are proven to be effective in practice. Model-Based Design streamlines development and enables automation. This facilities the execution of a lean process, increases the robustness of the processes, and leads to high-quality final software.
Lev Vitkin, Delphi Corporation
Decreasing time to market for advanced technologies requires alternative methodologies for efficient vehicle testing, calibration, and research. This session demonstrates the complete modeling of a vehicle with an electrical system for integrating fuel-saving technologies, including stop/start, coasting, and regenerative braking using MATLAB, Simulink, Simscape™, and Stateflow. A modified vehicle model as well as real-time execution, S-Function Builder, MATLAB scripts, and supporting hardware are used to display capabilities. It is evident that advanced engineering will include modeling and systems simulation in order to remain competitive in the future.
Data acquisition boxes are installed in Chrysler cars and trucks to measure their usage. The data is transmitted wirelessly and stored for future analysis. The resulting individual vehicle files are large, with about 20 million time-stamped records for 40 to 50 data channels collected at a rate of one and five readings per second for about a year. Traditionally, engineers have used tools like Microsoft® Excel® to analyze this data, but Excel is limited to 1,048,576 rows of data. This presentation discusses how Chrysler uses MATLAB to read and analyze the data and project customer usage.
Dennis Craggs, Chrysler Group LLC
Over the last decade, the Buckeye Bullet Program at The Ohio State University has set numerous international and national land speed records for electric vehicles, using both battery and fuel cell power systems. Designing and controlling these systems for a land speed racing application provides unique challenges that are not encountered in normal operation. This session presents an overview of the challenges and solutions for data analysis, modeling, and simulation of the 2009 and 2010 vehicles. Case studies of particularly interesting problems are covered, with a focus on the new 2013 400 mph vehicle.
R.J. Kromer, The Ohio State University Center for Automotive Research
Modern driver assistance systems implement the comprehension and judgment capabilities of a driver to interpret the events and observations on the road. This introduces a new form of complexity in the software to model the cognitive capabilities of a human. Traffic Sign Recognition and Display is an example of such a system. Observing the current traffic sign and vehicle maneuvers, retrieving the memory of what has been observed in the past, and considering the regional road and traffic rules all affect the logic behind selecting the valid traffic signs to display. This presentation outlines the creation of a configurable and reusable mathematical model in MATLAB and Simulink for logical interpretation of traffic signs and vehicle maneuvers. The benefits of using MATLAB and Simulink in the verification of this model are also discussed.
Reza Sahba, Magna Electronics
When working to improve vehicle energy efficiency, engineers must consider how the vehicle is operated. Different drivers and driving styles can create significant differences in actual fuel economy from the same type of vehicles. Eaton is developing LookAhead technology to minimize the drive odds for achieving the driving consistency for better fuel economy. The modularized system includes a GPS-based map database, mobile broadband traffic information, and on-board vehicular radar. With the application of Vehicle Network Toolbox™ and a single board computer, the LookAhead system can be integrated seamlessly into the Eaton UltraShift Plus powertrain system on a Class 8 long-haul truck. This presentation shows the details of system integration including related test data to demonstrate the effectiveness of this technology for improving the fuel economy.
Model-Based Design has become the preferred approach for developing automotive embedded software. This presentation describes the development of an engine setpoint controller using Model-Based Design and xPC Target™ Turnkey. The engine setpoint controller is designed to monitor engine rotation and other engine parameters and to control fuel, spark, and other engine actuators. The engine setpoint controller also allows control algorithms developed using Model-Based Design to be implemented and executed. Specific topics include the I/O driver designed with Simulink blocks, deployment of our custom VHDL® code to the I/O module of xPC Target, and a human machine interface design for the engine setpoint controller.
Sam Lee, General Motors Company
AUTOSAR is an emerging trend in the industry. As such, pilot programs for AUTOSAR must be executed alongside other production programs. Often, a model must be maintained for both AUTOSAR and non-AUTOSAR during the transition phase. This presentation demonstrates practical approaches for migrating a model’s architecture to improve the component-based decomposition of the system while maintaining the flexibility to support AUTOSAR and non-AUTOSAR programs. Use of models to facilitate integration of automatically generated and legacy handwritten code into an AUTOSAR framework are also discussed.
Dave Hoadley, MathWorks
Since the inception of the Japan MBD Automotive Advisory Board (JMAAB) more than 10 years ago, various aspects of verification have been addressed by multiple working groups. With the broadening adoption of Model-Based Development, model-based verification is considered critical for ensuring and improving quality. If implemented in the early stage, model-based verification will also help reduce development effort. The JMAAB Control Model Test Design Working Group was established to discuss practical methodologies to perform model verification. This presentation provides an overview of the work done by this working group, including test methods, test automation utilities, and mapping of the test methods to industry standards.
Yasumitsu Ito, MathWorks
Complying with the ISO 26262 functional safety standard requires a level of rigor in verification and validation, traceability, documentation, and tool selection that is higher than the current practice in the automotive industry. Assessing ISO 26262 compliance needs to take into consideration key attributes of Model-Based Design such as executable specification, model-based verification, and automatic code generation. This presentation introduces a framework for gap analysis, as well as a case study conducted using the framework.
Eric Dillaber, MathWorks
In this presentation, MathWorks engineers will present a process for developing lithium battery models for system level simulation and Battery Management System design. In this process, a battery cell model is built based on the electrochemical and thermal fingerprints of the battery. Nonlinearities are captured using look-up tables in Simscape™. Simulink Design Optimization™ is used to automate the parameterization process. As a result, simulation results are fitted to experimental discharge data. The model is then extended to a battery pack, taking advantage of the modularity of Simscape and the semiconductor element blocks in SimElectronics for external circuitry such as those for cell balancing. Finally, the model is optimized for speed which is essential for system level optimization and hardware in-the-loop testing.
Javier Gazzarri, MathWorks
Active safety features such as lane-departure warning and lane-keeping assist require complex algorithms. These algorithms process data from different sensors and control logic that determines the appropriate action to avoid an impending collision. Testing active safety features using a prototype vehicle can be difficult and even dangerous. This presentation introduces a tool chain for developing active safety features that supports the design and prototyping of computer vision algorithms, integrating computer vision algorithms with controller logic, and analyzing system performance, all within the simulation environment.
Chris Fillyaw, MathWorks
Real-world data never looks like clean waveforms as seen in the textbooks. Automotive engineers need to analyze large sets of test data from prototype vehicles, test cells, and databases, whether they work on powertrain, battery, vehicle dynamics, or durability. This session introduces techniques for dealing with this data and demonstrates how MATLAB and relevant toolboxes can be used to quickly and effectively:
Sumit Tandon, MathWorks
This master class uses a field-oriented control algorithm for a permanent magnet synchronous motor to illustrate how to generate efficient fixed-point C code from the controller model, integrate it with hand code for the embedded device drivers, and use the fully integrated software to spin the motor hardware. Topics include model architecture, algorithm export, scheduling techniques, code profiling, and code verification using processor-in-the-loop (PIL) testing.
Greg Wolff, MathWorks
Simscape and its libraries enable engineers to describe multidomain complex dynamic systems in a single environment, connecting blocks via physical connections, as opposed to the traditional Simulink signals. This presentation outlines the fundamentals of using MATLAB and Simulink for physical modeling, converging to an example that shows the design and optimization of a dual-clutch transmission system using SimDriveline™ in combination with Optimization Toolbox™ and Parallel Computing Toolbox™.
Vinay Gunasekaran, MathWorks