Academic Company Events NI Developer Zone Support Solutions Products & Services Contact NI MyNI

Customer Solutions

Honda Implements Integrated Road Simulator at Low Cost

Author(s):

Hiroshi Takahashi, Honda

Industry:

Automotive

Product:

LabVIEW, PXI/CompactPCI

The Challenge:

Developing a road simulator (RS) and virtual road simulator (VRS) for the designing and durability testing of Honda motorcycles.

The Solution:

Using LabVIEW 6 i, we developed a common control program for RS and VRS achieved by synchronizing a National Instruments analog output board and analog input board using a RTSI timing and triggering bus.


Staying Ahead of the Competition
Motorcycle development requires substantial technological innovation as competition intensifies worldwide. In our strength and durability test division, development has conventionally followed a cyclical, sequential method of prototyping, testing, and finally adopting a design. We needed a development approach capable of using computer simulation in the design phase and conducting simultaneous durability tests to meet the demand for a shorter development time period and fewer expensive preproduction vehicles. Road simulators (RS), a key tool for testing motorcycle strength and durability, are an important part of this new development approach.

We wanted to replace the current RS control system with a Windows NT-based system. The current MS-DOS-based RS control system did not provide satisfactory performance and was difficult to maintain.
We also wanted to use common control software for the RS and virtual road simulators (VRS). The validity of the VRS simulation model must be verified so there is no discrepancy between the results obtained by the RS and VRS (the data I/O interface portion of each program is not shared because the stimulus for the RS is physical while the stimulus for the VRS is virtual).

System Configuration of a Road Simulator
The RS control system consists of a PC, an analog output board (PCI-6711), an analog input board (PCI-6032E), and a signal conditioner. We simulate the rough road load conditions using an iterative process. First, we record the three desired responses (rear axle vertical acceleration, front axle vertical acceleration, and front fork longitudinal bending stress) during actual running on a rough road. We then install the test vehicle on the road simulator, activate each actuator with random noise, and measure the three responses to determine the transfer function matrix of the system. We execute trial activation using the obtained drive signals, measure the response signals, and calculate the deviation (error) between the desired responses and the achieved responses. The resulting error signals are convolved with the inverse of the transfer function matrix to create drive corrections that are added to the last drive signals.

Besides reducing the software budget, this low-cost system was also successful because we implemented a synchronous operation of activation and measurement of achieved responses. Therefore, it was not necessary to purchase an expensive dedicated I/O unit. At the start of the project, we were unsure if Windows NT would provide a stable input and output operation. We became more confident of achieving our goals by attending a National Instruments DAQ training course and receiving support on case examples of PCI bus product applications from field engineers.

Control of a Virtual Road Simulator
The software developed with LabVIEW controls three virtual actuators. Fatigue analysis software evaluates the result of activation, enabling us to predict the stress concentration "hot spots" on the vehicle and their stress waveform and fatigue life. Thus, we can now perform tests equivalent to RS tests on a PC. The software development involved the following key topics:

  • The displacement data to activate the virtual actuators is provided in binary file format to the Dynamic Analysis and Design System (DADS) model. The achieved responses calculated by DADS are also received in the same binary file format.
  • Actual run measurement uses a digital data recorder with a built-in, anti-aliasing filter of eighth-order Chebyshev function. Therefore, we must apply a filter of the same characteristics as the digital data recorder to the achieved responses calculated on the DADS side. Without the filtering, we cannot correctly perform evaluation when comparing the calculated result and measurement result. We used a LabVIEW IIR filter to create this functionality.
  • We implemented unmanned iteration functionality by using batch processing. This allows the operator to run the calculation overnight or do other jobs during calculation of the VRS iteration. Unmanned iteration was critical for successful implementation of a practical VRS system, and this was easily achieved by using the System Exec.vi and Get Command Line.vi downloaded from the NI Web site.
  • We developed the filtering and unmanned iteration in collaboration with LMS-CAE (who developed DADS) in the United States. By installing a LabVIEW runtime engine in the partner’s PC, we greatly reduced the size of the execution file, thus facilitating e-mail correspondence. Both the RS and the VRS can reproduce the load from actual run with simulation accuracy sufficient for practical use.

Saving Time and Money
Conventionally, RS control systems are large-scale and expensive. We also found that purchasing a complete control system for a single RS from an overseas manufacturer would cost approximately 35 million yen and approximately 10 million yen for a single copy of VRS control software. We are currently planning to update six sets of RS control systems. In addition, a substantial number of control software programs will be necessary after the VRS rolls out. The prohibitive cost for all this led us to in-house development. By developing the VRS control software using LabVIEW, we created the entire control system for less than one-tenth the cost of purchasing it from an overseas manufacturer (approximately 45 million yen). The total time taken to develop the control software for both the RS and VRS was approximately 220 days.
Honda R&D used LabVIEW to develop two types of road simulators for strength and durability tests of motorcycles - one for test benches and the other for cyber space. As a result, we are confident that we will achieve our goals of shortening motorcycle development time and reducing the number of expensive pre-production vehicles.

For more information, contact:

Shokichi Harashima or Hiroshi Takahashi

Honda R&D Co., Ltd.

Tel: 81 48 462 3300

E-mail: shokichi.harashima@mail.a.rd.honda.co.jp

View the PDF
shokichi_harashima_pc_2002.pdf

View the entire user solution in Adobe Acrobat PDF format.