Section 7.1 introduces welfare engineering research related to gait. Gait rehabilitation using biofeedback has shown high training efficacy for patients with neurological disorders. This method involves measuring biosignals such as Electromyography (EMG) and electroencephalography (EEG) and then using that information to consciously adjust one’s own movements. These approaches allow for the noninvasive creation of individualized training content. Gait training utilizing virtual reality (VR) and robotics makes it possible to set up diverse virtual environments and difficulty levels. In most cases, parameters like walking speed and sway can be quantified, offering advantages not found in conventional gait training. Furthermore, by combining biofeedback with robot-assisted rehabilitation, even higher intervention effects than traditional methods can be expected. In recent years, the development of various devices has enabled training that can intervene in both postural control and cognitive function. For example, dual-task training, such as “thinking while walking,” can enhance attention-splitting abilities, leading to improvements in cognitive and physical function, as well as promoting neuroplasticity (the ability to reorganize neural pathways). Artificial intelligence (AI)-powered pose estimation technology allows for low-cost and simple motion capture. As AI learning progresses, its measurement accuracy is also improving dramatically. This study addresses the growing need for mobility support in aging societies by developing and validating two assistive systems with real-time alert functions. The first system is a personal mobility (PM) system that uses IMU sensors to monitor body posture and detect misalignments between the user and the device. When abnormal joint angle deviations occur, such as between the trunk and backrest, vibration and voice alerts are triggered. Trials involving healthy adults simulating elderly conditions demonstrated that the alerts were activated within 0.3 s and perceived as effective. The second system is a walking-assist robot that uses HAL® and curara® exoskeletons. IMU sensors track hip and knee motion; deviations from preset angles trigger haptic alerts. The curara® system further synchronized these alerts with its assistive motions, resulting in smoother, more stable gait patterns. Compared to the HAL® system, the curara® system showed smaller joint deviations and higher user comfort. Although they are designed for different environments—the PM system for daily mobility and the walking-assist robot system for rehabilitation—both systems share a common architecture involving real-time sensing and multimodal feedback. This study emphasizes the importance of integrating these technologies into a unified support platform that can adapt to the user’s context and physical condition. Future development should combine PM and robotic assistance with AI and the Internet of Things (IoT), enabling personalized, adaptive mobility support that promotes safety and independent living for the elderly and disabled.