I. Introduction

The electrification of transport is accelerating as governments and manufacturers seek ways to reduce greenhouse gas emissions and local pollutants from road transport. Battery electric vehicles (BEVs) are central to this transition; however, their effective deployment, from powertrain design to regenerative braking policies and eco-driving strategies, relies on modelling tools that are physically accurate and computationally efficient. Standardised procedure (e.g., Worldwide harmonized Light vehicles Test Procedure (WLTP) and Worldwide harmonized Light-duty vehicles Test Cycle (WLTC)) provide a common benchmark, but real-world operation often differs from laboratory cycles due to factors such as traffic, inclines, environmental conditions and additional loads, which makes it necessary to develop simulators that are based on fundamental principles and calibrated using real-world data[1]. Real-world energy estimation studies further refine prediction under traffic and gradient variability [2]–[4]. A rich literature has addressed EV energy prediction across levels of abstraction. Longitudinal models can be categorised broadly as either backward (powerflow) or forward (driver-in-the-loop) approaches. The latter approach explicitly captures tracking dynamics, actuator limits and transient effects at a modest computational cost [5], [6]. Widely-used tool chains such as MATLAB/Simulink and established platforms (e.g., FASTSim) enable rapid development and scenario studies [7]. Power-based and map-based formulations have demonstrated strong accuracy when properly parameterized and validated against measurements [2] [4]. On the traction side, the permanent-magnet synchronous motor (PMSM) and inverter determine conversion efficiency. Field-weakening extends the constant-power region, while efficiency-map or loss-model representations help capture speed torque dependent losses across the operating envelope [8]. Regenerative braking further couples motor/inverter and battery constraints with tire adhesion and stability considerations; recent strategies target efficiency while preserving safety and drivability, especially under low-speed or low- conditions [9], [10]. Battery State-of-Charge (SOC) estimation is equally critical. While Coulomb counting is the baseline for vehicle-level energy accounting, accuracy improves when combined with equivalent-circuit models, temperature effects, and stochastic filters. Foundational texts and studies compare extended/unscented Kalman filters and more recent adaptive or cubature variants, as well as data-driven and deep-learning approaches that account for nonlinearity and hysteresis [11] [16]. Despite the progress, many SOC estimators are validated at cell/module scale with high-rate laboratory data, whereas vehicle-level forward simulations often must rely on coarse telematics sampling. Three gaps persist in the state of the art. (i) End-to-end vehicle-level validation against production EV data collected on public roads is comparatively scarce, particularly with explicit error metrics on both energy and SOC when only coarse (tens-of-seconds) samples are available [2], [3]. (ii) The traction chain is sometimes simplified via scalar efficiencies; fewer studies model and validate, as a single integrated system, the PMSM torque-speed envelope, inverter constraints, field-weakening, and regenerative braking limits for a specific drivetrain layout (e.g., rear-wheel drive). (iii) Sensitivity analyses that disentangle the contributions of aerodynamic/rolling parameters, internal resistance, auxiliary loads, and regeneration policy to route-level energy remain limited in vehicle-specific, real-route contexts [4], [5], [7]. We develop a modular, forward-facing longitudinal simulator for a commercial BEV the Tesla Model-Y in Rear-Wheel-Drive (RWD) configuration implemented in MATLAB/Simulink and experimentally validated on an urban route recorded on-road. The model comprises Driver, Brake System with regenerative blending, PMSM Motor/Inverter with efficiency map and field-weakening, Driveline, Battery (Thevenin + Open Circuit Voltage-SOC) with Coulomb counting, and Longitudinal Dynamics. Calibration uses measured/nominal vehicle parameters; validation reports Mean Absolute Percentage Error (MAPE) for energy and SOC, achieving over the route. Relative to prior art, our contributions are: (i) a transparent, re-usable Simulink architecture whose constraints and maps are documented for replication; (ii) an integrated PMSM/inverter/regeneration chain consistent with RWD limits and coarse-sampling validation and (iii) a sensitivity analysis clarifying the dominant levers on route energy (aerodynamics, rolling, internal resistance, auxiliaries, regeneration), situating results alongside standardized-cycle [1]and prior energymodeling baselines [2], [5]. The article's structure is outlined as follows: Section II details the model components (dynamics, PMSM/inverter with field-weakening, battery/SOC, driveline) and the validation protocol. Section III presents results on speed/SOC tracking, aggregate energy accuracy, and sensitivity. Section IV discusses limitations and use cases. Section V concludes and outlines future work (thermal coupling, hybrid SOC estimators, multi-motor wheel drive), extending the modeling fidelity and generality suggested by recent reviews.