Zahid Anwar
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Zahid Anwar has a total of 3 co-authors. Here are the top co-authors by number of shared publications:
- Muhammad Zubair Malik
Computer Science, North Dakota State University, Fargo, United States, USA
1 shared publication - Nasir Ahtasham
Computer Science, North Dakota State University, Fargo, United States, USA
1 shared publication - Zia Muhammad
Department of Computing, Design, and Communication, University of Jamestown, Jamestown, ND 58405, USA
1 shared publication
Latest Publications
- Open Access
- 18 November 2026
- Published by MDPI AG in Preprints.org
RLOCS: Real-Time Operational Mode Advisement and Temporal Command Acceptance Using Standardized OBD-II Signals
Drive-mode advisement requires both reliable classification and explicit rules for accepting and communicating recommendations. This paper presents a modular framework that links longitudinal vehicle signals to Eco, Tour, and Sport advisement, temporal mode acceptance, and simulated subsystem-command generation. The input design uses throttle position, engine speed, vehicle speed, and longitudinal acceleration, with an optional engineered throttle–engine-speed interaction. Logistic regression, nearest neighbors, and random forest are compared on 67,440 recorded samples using a fixed stratified random-row split and training-only preprocessing. The four-input random forest achieves 96.98% accuracy and a macro-averaged F1 score of 0.9683; the five-input reference achieves 96.81% accuracy. Controlled feature ablation and 180 throttle-perturbation runs characterize representation dependence and input sensitivity. The compiled five-input reference has a mean single-request host latency of 43.08 microseconds and 30.890 mebibytes of static model data. A MATLAB/Simulink implementation combines confidence gating, a dwell-time state machine, parameter lookup, command bounds, slew limits, and software controller-area-network encoding. Five baseline scenarios and a 45-run sensitivity sweep reproduce configured acceptance delays, suppress brief erroneous recommendations, and preserve command consistency. The results establish classification within the available recordings and an executed software advisory-to-command path.
- 18 November 2026
- Published by MDPI AG in Preprints.org
Open AccessRLOCS: Real-Time Operational Mode Advisement and Temporal Command Acceptance Using Standardized OBD-II Signals
Drive-mode advisement requires both reliable classification and explicit rules for accepting and communicating recommendations. This paper presents a modular framework that links longitudinal vehicle signals to Eco, Tour, and Sport advisement, temporal mode acceptance, and simulated subsystem-command generation. The input design uses throttle position, engine speed, vehicle speed, and longitudinal acceleration, with an optional engineered throttle–engine-speed interaction. Logistic regression, nearest neighbors, and random forest are compared on 67,440 recorded samples using a fixed stratified random-row split and training-only preprocessing. The four-input random forest achieves 96.98% accuracy and a macro-averaged F1 score of 0.9683; the five-input reference achieves 96.81% accuracy. Controlled feature ablation and 180 throttle-perturbation runs characterize representation dependence and input sensitivity. The compiled five-input reference has a mean single-request host latency of 43.08 microseconds and 30.890 mebibytes of static model data. A MATLAB/Simulink implementation combines confidence gating, a dwell-time state machine, parameter lookup, command bounds, slew limits, and software controller-area-network encoding. Five baseline scenarios and a 45-run sensitivity sweep reproduce configured acceptance delays, suppress brief erroneous recommendations, and preserve command consistency. The results establish classification within the available recordings and an executed software advisory-to-command path.