Max Robotics — Global Service Robot Directory + US Certification

About

Gabriele Costante is a robotics and computer vision researcher whose work spans autonomous navigation, visual perception, and deep learning-based scene understanding. His research addresses some of the most pressing challenges in mobile robotics, including monocular depth estimation, visual odometry, and target-driven navigation. Costante's most influential contributions lie in leveraging deep learning to solve traditionally geometry-dependent problems. His 2016 work on fast monocular depth estimation for obstacle detection using fully convolutional networks (105 citations) demonstrated that robust, real-time depth perception was achievable without specialized hardware—critical for high-speed autonomous systems. Complementing this, his work on domain-independent depth estimation (69 citations) pushed toward more generalizable learning-based approaches. His investigations into data-driven visual odometry, including uncertainty estimation (49 citations), have further strengthened the reliability of learning-based localization methods. Beyond perception, Costante has made notable contributions to visual navigation through deep reinforcement learning (102 citations), perception-aware path planning (37 citations), and even agricultural robotics through unsupervised fruit counting (38 citations). His work on natural language interfaces for service robots reflects a broader vision of integrated, human-aware robotic systems. With over 570 total citations across diverse topics, Costante stands as a versatile and impactful figure in modern autonomous robotics research.