Electrical Engineering Approaches for AI-Enabled Autonomous Robotics in U.S. Industrial and Defense Applications

Abstract

This quantitative study examined how electrical engineering approaches influenced the performance of AI-enabled autonomous robotics in U.S. industrial and defense applications. A total of 220 valid responses were analyzed, including 112 industrial (50.9%) and 108 defense (49.1%) system cases, ensuring balanced cross-domain representation. Six electrical engineering constructs were evaluated as predictors of autonomy performance: sensor system performance, embedded computing efficiency, power electronics and energy stability, control and actuation accuracy, communication system reliability, and safety and reliability engineering effectiveness. Descriptive results indicated that safety and reliability engineering demonstrated the highest consistency (M = 4.12, SD = 0.49), followed by control and actuation accuracy (M = 3.94, SD = 0.55) and sensor system performance (M = 3.82, SD = 0.61). Embedded computing efficiency (M = 3.56, SD = 0.74) and power stability (M = 3.41, SD = 0.79) showed the greatest variability across systems. Reliability analysis confirmed strong internal consistency across constructs, with Cronbach’s alpha coefficients ranging from 0.842 to 0.926. Multiple regression analysis demonstrated that the combined model significantly predicted autonomy performance (R² = 0.659, Adjusted R² = 0.647, F = 67.83, p < .001). Embedded computing efficiency (β = 0.338, p < .001) and sensor system performance (β = 0.301, p < .001) were the strongest predictors. Power stability (β = 0.256, p < .001) and control accuracy (β = 0.213, p < .001) also significantly contributed. Communication reliability showed a moderate effect (β = 0.158, p = .001) and was stronger in defense contexts, while safety engineering provided a smaller but significant stabilizing influence (β = 0.109, p = .014). Findings confirmed that electrical subsystem performance significantly explained variation in AI-enabled autonomy outcomes across both domains.