Scholarlink

Research has suggested that systems engineeringSystems engineering methods may need to be evolved to effectively support the engineering of intelligent systemsIntelligent systems. This paper presents the prototype of a simulationSimulation testbedTestbed that enables the evaluation of systems engineeringSystems engineering methods for intelligent systemsIntelligent systems. The testbedTestbed is a surrogate of an intelligent detection system composed of an optical lens, a camera, and a classification system. While the intelligent algorithm is implemented as part of the classification system, the results of intelligence are a function of the performance of the lens and the camera as well. The models of the three components are designed to enable modifications that represent those that an actual optical detection system may exhibit in real life, including manufacturing variations (such as lens transmission or pixel defects), technology upgrades or obsolescence (such as increase of camera resolution), or operational wear (such as lens transmission deterioration or pixel cracks). We show in this paper that such a design allows for (1) inducing variations that a real system would exhibit and (2) such variations enable the assessment of the effectiveness of systems engineeringSystems engineering methods for intelligent systemsIntelligent systems.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A Prototype Simulation Testbed to Evaluate Systems Engineering Methods for Intelligent Systems

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Samuel Cornejo,
  • Alejandro Salado

摘要

Research has suggested that systems engineeringSystems engineering methods may need to be evolved to effectively support the engineering of intelligent systemsIntelligent systems. This paper presents the prototype of a simulationSimulation testbedTestbed that enables the evaluation of systems engineeringSystems engineering methods for intelligent systemsIntelligent systems. The testbedTestbed is a surrogate of an intelligent detection system composed of an optical lens, a camera, and a classification system. While the intelligent algorithm is implemented as part of the classification system, the results of intelligence are a function of the performance of the lens and the camera as well. The models of the three components are designed to enable modifications that represent those that an actual optical detection system may exhibit in real life, including manufacturing variations (such as lens transmission or pixel defects), technology upgrades or obsolescence (such as increase of camera resolution), or operational wear (such as lens transmission deterioration or pixel cracks). We show in this paper that such a design allows for (1) inducing variations that a real system would exhibit and (2) such variations enable the assessment of the effectiveness of systems engineeringSystems engineering methods for intelligent systemsIntelligent systems.