This paper proposes a shared control support system for electric wheelchairs, aiming to enhance the control experience for individuals with mobility impairments. Traditional joystick control requires dexterous hand movements and high cognitive ability, making it difficult for elderly users with motor or cognitive impairments. Many existing shared control systems focus on obstacle avoidance, but these approaches perform poorly in environments with limited clearance, such as door gaps. To address this, we introduce a system that predicts the user’s intention based on past wheelchair trajectories, without the need for prior environmental information. The system employs a gap-based local path planner that detects immediate surrounding gaps using a 2D LIDAR sensor, predicting the intended gap of the user and guiding the wheelchair through it. To prevent false positives, a “No Gap” hypothesis is used when the user do not intend to go through any of the found gaps. This approach ensures the system provides support that aligns with the user’s intention, improving control in tight or clustered environments.