Abstract:Enabling robots to autonomously navigate unknown, complex, dynamic environments and perform diverse tasks remains a fundamental challenge in developing robust autonomous physical agents. They must effectively perceive their surroundings while leveraging world knowledge for decision-making. While recent approaches utilize vision-language and large language models for scene understanding and planning, they often rely on offline processing, external computing, or restrictive environmental assumptions. We present a novel framework for efficient and scalable real-time, onboard autonomous navigation that integrates multi-level abstraction in both perception and planning in unknown large-scale environments that change over time. Our system fuses data from multiple onboard sensors for localization and mapping and integrates it with open-vocabulary semantics to generate hierarchical scene graphs. An LLM-based planner leverages these graphs to generate high-level task execution strategies, which guide low-level controllers in safely accomplishing goals. Our framework's real-time operation enables continuous updates to scene graphs and plans, allowing swift responses to environmental changes and on-the-fly error correction. This is a key advantage over static or rule-based planning systems. We demonstrate our system's efficacy on a quadruped robot navigating large-scale, dynamic environments, showcasing its adaptability and robustness in diverse scenarios.
Abstract:Large-scale LiDAR mappings and localization leverage place recognition techniques to mitigate odometry drifts, ensuring accurate mapping. These techniques utilize scene representations from LiDAR point clouds to identify previously visited sites within a database. Local descriptors, assigned to each point within a point cloud, are aggregated to form a scene representation for the point cloud. These descriptors are also used to re-rank the retrieved point clouds based on geometric fitness scores. We propose SALSA, a novel, lightweight, and efficient framework for LiDAR place recognition. It consists of a Sphereformer backbone that uses radial window attention to enable information aggregation for sparse distant points, an adaptive self-attention layer to pool local descriptors into tokens, and a multi-layer-perceptron Mixer layer for aggregating the tokens to generate a scene descriptor. The proposed framework outperforms existing methods on various LiDAR place recognition datasets in terms of both retrieval and metric localization while operating in real-time.
Abstract:Most of the deep learning based speech enhancement (SE) methods rely on estimating the magnitude spectrum of the clean speech signal from the observed noisy speech signal, either by magnitude spectral masking or regression. These methods reuse the noisy phase while synthesizing the time-domain waveform from the estimated magnitude spectrum. However, there have been recent works highlighting the importance of phase in SE. There was an attempt to estimate the complex ratio mask taking phase into account using complex-valued feed-forward neural network (FFNN). But FFNNs cannot capture the sequential information essential for phase estimation. In this work, we propose a realisation of complex-valued long short-term memory (RCLSTM) network to estimate the complex ratio mask (CRM) using sequential information along time. The proposed RCLSTM is designed to process the complex-valued sequences using complex arithmetic, and hence it preserves the dependencies between the real and imaginary parts of CRM and thereby the phase. The proposed method is evaluated on the noisy speech mixtures formed from the Voice-Bank corpus and DEMAND database. When compared to real value based masking methods, the proposed RCLSTM improves over them in several objective measures including perceptual evaluation of speech quality (PESQ), in which it improves by over 4.3%