Description
This Code attempts to aid military vehicle system representatives with design specific test strategies that have an increased relevance to current actual operating conditions while taking into account the PM’s and the test center’s abilities and constraints.
A test’s relevance often forms trade-offs with time, effort, and cost.
By employing mathematical optimization techniques, exploring the wide array of options open to the PM, and weighing needs and desires against available data, time, and staffing, the PM or customer can more confidently choose the most relevant vehicle test plan that best suites his requirements and his constraints.
The rapid deployment of automotive systems has caused the Department of Defense test community and the Aberdeen Test Center in particular to reevaluate and redefine traditional test plans and practices in order to maximize the amount of valid and pertinent data obtained from shortened test schedules. As a result, this thesis studies new transformation plans to provide ways to optimize military test plans.
These transformation plans take into account existing military vehicle data from multiple sources including the Aberdeen Test Center’s automotive road courses. These transformation plans are not only useful for shortened military tests, but can also be easily employed in developing test plans for private industry customers as well as long term test projects. The benefits in all cases are the same: an optimized test plan for automotive endurance operations.
The Problem and a Solution
The rapid deployment of automotive systems has caused the Department of Defense test community and the Aberdeen Test Center in particular to reevaluate and redefine traditional test plans and practices in order to maximize the amount of valid and pertinent data obtained from shortened test schedules. However, the process of creating a detailed test plan can require significant time and effort. This process is called a transformation plan because it transforms information about customer requirements and operational data into a detailed test plan. ATC customers desire transformation plans that create highly relevant detailed test plans using the least amount of time and cost.
Unfortunately, test planning can become routine. Development programs simply reuse the detailed test plan that was used last time without investigating its relevance to new environments. This routine transformation plan reduces the time and effort involved but can lead to inappropriate test plans. Consider, for instance, using a test plan developed soon after World War II for testing trucks that, sixty years later, will be driving not on the streets of European cities but instead on highways through Middle Eastern countryside.
To address this problem, a set of transformation plans were systematically developed that can be used to create highly relevant detailed test plans using the least amount of time and cost. The relative performance of these plans was evaluated as they were implemented for two common military vehicle systems. This work studied two specific cases: the M915 truck tractor and the M1114 HMMWV. A set of feasible transformation plans that are relevant to both cases were created. Both vehicle systems were soon to undergo automotive endurance testing at ATC, and, for both vehicles, data obtained from 30 days of operation in OIF was available in the EUDB.
The data available for each vehicle system was used in an optimization algorithm developed for two of the transformation plans. The optimization algorithm was used to mathematically determine optimized road course test matrices using both single and multi-objective optimization techniques by minimizing the L2 Norm error between the actual use data captured from OIF and ATC road course data for both vehicles. The M915 truck tractor road course test matrix was first optimized for a single channel, road speed, using Transformation Plan C-1, and then was optimized in combination with a second data channel, transmission temperature, in Transformation Plan C-2. In similar fashion, the M1114 HMMWV road course test matrix was optimized first for road speed relevance, and then in combination with the roll rate data channel for its two optimization transformation plans.
A third set of road course test matrices was created from Transformation Plan A for both vehicles. The transformation plans were then compared on the following objective and subjective performance metrics: the fit between the test plan and the operational conditions (measured using a correlation coefficient and an error measurement), the cost to implement the transformation plan, the computational effort, the time to execute the transformation plan, the expertise required, the effort required, and the customer’s satisfaction with the results.
This code has a document (194 pages) which describe the algorithm in detail.
Selected outputs :
https://matlab1.com/wavelet-transformation/
https://isr.umd.edu/~jwh2/papers/IEST.pdf
Reviews
There are no reviews yet.