Search Results

Model based design has proven itself as a key enabler to accelerate the development and enhance the quality of embedded software during the last decade. The tools serving the model based design have continuously improved both in terms of software and hardware, and have achieved a maturity level capable to provide seamless transition across different development phases. This approach provides a capability to ensure the satisfaction of key requirements not only at controls development level but also at system level directly addressing the customer requirements. This paper demonstrates a modeling architecture utilizing this capability, allowing the user work on the complete system model that is capable to maintain the same model architecture starting from pure model simulation phase working its way through vehicle testing, contributing to the reduction of the development time and system integration effort while improving the utilization of the model based tools and methods.

A patent pending engine control system with torque sensor feedback is described. Upon detecting a loss in traction by means of a torque sensor, engine torque is adjusted via throttle paving the way for improved traction and enhanced stability. The throttle is reduced to a calculated value using engine characteristics, the torque sensor measurement and non-slipping wheel speed information. The advantages of the powertrain torque sensing as opposed to speed sensing are demonstrated thru a case study of a RWD SUV with an open rear differential. Simulations are used to prove the concept while the bandwidths of a number of physical systems contributing to the overall response time are ignored. Therefore the data provided in this paper should be treated relatively comparing speed sensors versus torque sensors. There are a number of engine torque reduction methods faster than throttle control such as spark retard and fuel shutoff.

As the active chassis technology becomes more and more sophisticated, it becomes increasingly important to have a repeatable and objective environment for testing to accelerate the deployment of active chassis control systems. Track testing is a highly subjective and expensive method for testing and validation of active chassis control algorithms. It is often difficult to recreate a critical condition that a vehicle chassis experiences within its possible operation range during track testing. Therefore an environment with capability to create a critical condition in a repeatable and objective manner is highly desirable. This paper presents a repeatable and objective method for developing as well as measuring performance of active chassis systems.

This paper presents a uniquely integrated environment where a control system is modeled, simulated and implemented in real time requiring no coding, no compiling nor embedding for multiple test rigs. The novel hardware architecture simultaneously configures itself according to the contents of a visual configuration editor (VCE) that is put together by the user utilizing graphical representations of basic building blocks called primitives. Similar to an analog computer, as the user builds the VCE model, the “code is simultaneously compiled and embedded” into the target DSP processor/s as easy as dragging and dropping primitives from various libraries into the VCE and connecting them. Similarly DSP processors within the hardware network are visually available to the user to be dropped into the VCE creating a distributed multi-processor control environment. The execution of multiple DSPs take place simultaneously reaching 25.6 kHz loop rates and 200 kz sampling rates.

This paper presents a novel method to measure a vehicle's rollover propensity. The measurement is conducted on a test rig consisting of a vehicle standing on a platform, dynamically loaded to initiate a potential rollover phenomenon. A set of specific tests provides repeatable quantitative results for a vehicle's rollover propensity. By testing a batch of vehicles of the same class, a relative rollover propensity-scoring table is established. The dynamic method is proposed as an alternative to NHTSA's quasi-static five star scoring system to measure vehicle's rollover propensity.

It is becoming a greater challenge to win a race as more and more technologies are put into use to enhance vehicle-handling characteristics. Vehicle handling during cornering and braking is one of the crucial performance characteristics that is continuously improved upon. To help the driver achieve the best handling performance on the track, suspensions are tuned to optimize for the specific track and racing conditions. Typically 4-post test rigs are used to test for the heave, pitch, roll and torsional frequencies of the suspensions. However three more actuators (aeroload actuators) become necessary to tune the suspensions at the correct ride height by including the aerodynamic effects. By means of a 7-post testing system an optimization procedure is executed to improve handling performance. An instrumented vehicle is driven on the track. The response of the vehicle is then replicated on a 7-post system using Iterative Control System (ICS) software.

Vehicle ride comfort and handling demands conflicting design requirements from a vehicle suspension system. As a result, a passive vehicle suspension is incapable of providing the optimum performance for both ride and stability. A test set up is presented in this paper to enable a developer to test any given control algorithm on real-time hardware for an active vehicle suspension, and simulate quarter car performance characteristics. A quarter car model is developed in a host computer and then embedded in a target computer. The parameters of the quarter car are accessible for modification in host computer through TCP/IP connection with target computer. Ride and stability performance characteristics are calculated and also made available during HIL simulation. A controller including a DSP card works with a third computer, where the control system architecture can be edited and control parameters can be changed using a customizable graphical user interface.

In this paper, a practical modeling method is presented, integrating simulation with test and measurement for a vibration table. Servovalve, hydraulic actuator and load are modeled as main components of a single channel vibration table. The system components are tested under both steady state and dynamic conditions. Servovalve steady state measurements are used to model nonlinear behavior instead of using purely parametric approach. Frequency response data is used in linear 2nd order transfer function approximation for the system. Mathematical models and look-up tables are generated using fundamental equations and conducting basic experiments. Developed models are later integrated to build the complete model for the vibration table in MATLAB-SIMULINK. The obtained model is intended to develop prototype models for any given vibration table configuration in a flexible and reliable design environment.