Autonomous vehicles require a precise control system capable of handling nonlinear vehicle dynamics safely perform agile maneuvers, such as lane changes. However, in real-world operating conditions, factors such as sensor measurement noise, process noise, time delays, and data packet loss can compromise the stability of the control system. In this research, an integrated control framework based on Nonlinear Model Predictive Control (NMPC) and Moving Horizon Estimation (MHE) is proposed. In the proposed method, the MHE filters out noise effects and reconstructs the system states during periods of data loss and time delays by incorporating physical constraints and the dynamic model. Subsequently, the NMPC receives the corrected states and calculates optimal commands aimed at minimizing the tracking error and maintaining passenger comfort. Simulation results of a double lane change maneuver at a speed of 108 km/h demonstrate that in a critical scenario (comprising a 100 ms network delay, 20% data packet loss, sensor noise, and uncertainties arising from employing a twin-track model with the Pacejka tire formula in the simulation plant), the proposed approach exhibits superior performance compared to the Extended Kalman Filter (EKF) algorithm. This structure fully maintains the vehicle's stability while preventing undesirable oscillations in the steering angle and traction force. Recording a maximum lateral error of 0.1 m and a Root Mean Square Error (RMSE) of 0.044 m demonstrates the outstanding performance of this system in ensuring safety and stability.