東芝情報システム Product Lineup

Are there any challenges related to "self-localization and environmental recognition (SLAM)" in manufacturing sites and autonomous systems? ● Unable to determine self-location in factory or indoor environments where GPS cannot be used ● Decreased localization accuracy due to environmental changes (lighting, layout changes) ● High-precision self-localization is computationally intensive and cannot operate in real-time ● Difficulties in integrating multiple sensors such as cameras, IMUs, and odometry 　　　　　　　　　　　　　▼ Leave it to our SLAM solution! ――― Features ―――――――――――――――――――――― ◆ Visual SLAM that operates with low resources ◆ Improved accuracy and robustness through sensor fusion ◆ Recognition of the surrounding environment using surrounding map generation functionality ――――――――――――――――――――――――――― SLAM is a technology that allows mobile entities (AGVs and autonomous systems) to map their surrounding environment while moving and to identify their own location. It observes features in the environment, estimates self-location based on that information, and simultaneously creates an environmental map. Our company primarily develops self-localization technology/SLAM using cameras.

In AGVs and autonomous robots, it is necessary to accurately grasp the destination and current position to determine the travel route. This solution customizes Toshiba's unique self-position estimation technology for your system and optimizes it for real-time systems. ■ Do you have any of these challenges? - Want to improve the self-position estimation accuracy of AGVs and transport robots - Want to reduce position drift and cumulative errors that occur during long-distance travel - Want to implement SLAM that is robust to environmental changes such as ambient light, lighting, and moving obstacles - Want to combine multiple sensors such as cameras, LiDAR, and IMUs - Want to achieve autonomous driving in areas where GPS/GNSS signals are unstable - Want to generate an environmental map while simultaneously estimating self-position - Need a fast, low-memory SLAM that operates on embedded devices - Want to customize SLAM to fit your own robots or mobile units ■ Would you like to solve these challenges with SLAM? We support self-position estimation tailored to your system by combining Visual SLAM, Visual Relocalizer, landmark-based SLAM, sensor fusion, and environmental map generation.