Accurate identification of complex terrain characteristics, such as soil composition and coefficient of friction, is essential for model-based planning and control of mobile robots in off-road environments. Spectral signatures leverage distinct patterns of light absorption and reflection to identify various materials, enabling precise characterization of their inherent properties. Recent research in robotics has explored the adoption of spectroscopy to enhance perception and interaction with environments. However, the significant cost and elaborate setup required for mounting these sensors present formidable barriers to widespread adoption. In this study, we introduce RS-Net (RGB to Spectral Network), a deep neural network architecture designed to map RGB images to corresponding spectral signatures. We illustrate how RS-Net can be synergistically combined with Co-Learning techniques for terrain property estimation. Initial results demonstrate the effectiveness of this approach in characterizing spectral signatures across an extensive off-road real-world dataset. These findings highlight the feasibility of terrain property estimation using only RGB cameras.

The thirteenth AAAI Mobile Robot Competition and Exhibition was once again collocated with AAAI-2204, in San Jose, California. As in previous years, the robot events drew competitors from both academia and industry to showcase state-ofthe- art mobile robot software and systems in four organized events.

Running services in many small processes improves fault tolerance since any number of services can fail due to programming faults without affecting the rest of the system. While it is clearly important to be able to handle a wide range of failures, application authors should not be required to implement routines to test and react in every known mode of failure for every application, even if the failures are abstracted to a common interface. Thus, the framework also provides transparent fault-tolerance to users of system services. Errors in software and hardware are detected, and corrective action is taken. Services can be restarted or removed from the system, and clients are reconnected to the same service or to another service implementing the same interface without intervention from the application programmer. The Washington University team successfully demonstrated its failure-tolerant framework on its robot, Lewis (figure 6).

In an attempt to solve as much of the AAAI Robot Challenge as possible, five research institutions representing academia, industry, and government integrated their research into a single robot named GRACE. This article describes this first-year effort by the GRACE team, including not only the various techniques each participant brought to GRACE but also the difficult integration effort itself.

In an attempt to solve as much of the AAAI Robot Challenge as possible, five research institutions representing academia, industry, and government integrated their research into a single robot named GRACE. This article describes this first-year effort by the GRACE team, including not only the various techniques each participant brought to GRACE but also the difficult integration effort itself.