Abstract--This paper presents an Autonomous System (AS) architecture for vehicles in a closed circuit. The AS performs precision tasks including computer vision for environment perception, positioning and mapping for accurate localization, path planning for optimal trajectory generation, and control for precise vehicle actuation. Each subsystem operates independently while connecting data through a cohesive pipeline architecture. The system implements a modular design that combines state-of-the-art technologies for real-time autonomous navigation in controlled environments. Autonomous vehicle systems in controlled environments present significant challenges in integrating multiple subsystems for real-time navigation and decision-making. The development of modular architectures that effectively combine perception, localization, path planning, and control systems represents a critical area of research in autonomous driving technology. This work presents a comprehensive framework for the connectivity and allocation of responsibilities within an autonomous driving architecture, focusing on precise operation in closed-circuit scenarios. The approach defines four primary modules: perception, localization and mapping, trajectory planning, and control.

This paper introduces a novel Kalman filter framework designed to achieve robust state estimation under both process and measurement noise. Inspired by the Weighted Observation Likelihood Filter (WoLF), which provides robustness against measurement outliers, we applied generalized Bayesian approach to build a framework considering both process and measurement noise outliers.

Learning-based inertial odometry has achieved remarkable progress in pedestrian navigation. However, extending these methods to quadruped robots remains challenging due to their distinct and highly dynamic motion patterns. Models that perform well on pedestrian data often experience severe degradation when deployed on legged platforms. To tackle this challenge, we introduce X-IONet, a cross-platform inertial odometry framework that operates solely using a single Inertial Measurement Unit (IMU). X-IONet incorporates a rule-based expert selection module to classify motion platforms and route IMU sequences to platform-specific expert networks. The displacement prediction network features a dual-stage attention architecture that jointly models long-range temporal dependencies and inter-axis correlations, enabling accurate motion representation. It outputs both displacement and associated uncertainty, which are further fused through an Extended Kalman Filter (EKF) for robust state estimation. Extensive experiments on public pedestrian datasets and a self-collected quadruped robot dataset demonstrate that X-IONet achieves state-of-the-art performance, reducing Absolute Trajectory Error (ATE) by 14.3% and Relative Trajectory Error (RTE) by 11.4% on pedestrian data, and by 52.8% and 41.3% on quadruped robot data. These results highlight the effectiveness of X-IONet in advancing accurate and robust inertial navigation across both human and legged robot platforms.

The extended Kalman filter (EKF) has been the industry standard for state estimation problems over the past sixty years. The Invariant Extended Kalman Filter (IEKF) is a recent development of the EKF for the class of group-affine systems on Lie groups that has shown superior performance for inertial navigation problems. The IEKF comes in two versions, left- and right- handed respectively, and there is a perception in the robotics community that these filters are different and one should choose the handedness of the IEKF to match handedness of the measurement model for a given filtering problem. In this paper, we revisit these algorithms and demonstrate that the left- and right- IEKF algorithms (with reset step) are identical, that is, the choice of the handedness does not affect the IEKF's performance when the reset step is properly implemented. The reset step was not originally proposed as part of the IEKF, however, we provide simulations to show that the reset step improves asymptotic performance of all versions of the the filter, and should be included in all high performance algorithms. The GNSS-aided inertial navigation system (INS) is used as a motivating example to demonstrate the equivalence of the two filters.

State estimation is crucial for legged robots as it directly affects control performance and locomotion stability. In this paper, we propose an Adaptive Invariant Extended Kalman Filter to improve proprioceptive state estimation for legged robots. The proposed method adaptively adjusts the noise level of the contact foot model based on online covariance estimation, leading to improved state estimation under varying contact conditions. It effectively handles small slips that traditional slip rejection fails to address, as overly sensitive slip rejection settings risk causing filter divergence. Our approach employs a contact detection algorithm instead of contact sensors, reducing the reliance on additional hardware. The proposed method is validated through real-world experiments on the quadruped robot LeoQuad, demonstrating enhanced state estimation performance in dynamic locomotion scenarios.

Muhammad Hafil Nugraha Research Centre for Smart Mechatronics National Research and Innovation Agency Bandung, Indonesia muha167@brin.go.id Estiko Rijanto Research Centre for Smart Mechatronics National Research and Innovation Agency Bandung, Indonesia estiko.rijanto@brin.go.id Oka Mahendra Research Centre for Smart Mechatronics National Research and Innovation Agency Bandung, Indonesia oka.mahendra@brin.go.id Abstract -- Accurate localization is crucial for effectively operating mobile robots in indoor environments. This paper presents a comprehensive approach to mobile robot localization by integrating an ultrasound - based indoor positioning system (IPS) with wheel odometry data via sensor fusion techniques. The Extended Kalman Filter (EKF) fusion method combines the data from the IPS sensors and the robot's wheel odometry, providing a robust and relia ble localization solution. Extensive experiments in a controlled indoor environment reveal that the fusion - based localization system significantly enhances accuracy and precision compared to standalone systems.

--This paper presents the design of a pose estimator for a four wheel independent steer four wheel independent drive (4WIS4WID) wall climbing mobile robot, based on the fusion of multimodal measurements, including wheel odometry, visual odometry, and an inertial measurement unit (IMU) data using Extended Kalman Filter (EKF) and Unscented Kalman Filter (UKF). The pose estimator is a critical component of wall climbing mobile robots, as their operational environment involves carrying precise measurement equipment and maintenance tools in construction, requiring information about pose on the building at the time of measurement. Due to the complex geometry and material properties of building fac ades, the use of traditional localization sensors such as laser, ultrasonic, or radar is often infeasible for wall-climbing robots. Moreover, GPS-based localization is generally unreliable in these environments because of signal degradation caused by reinforced concrete and electromagnetic interference. Consequently, robot odometry remains the primary source of velocity and position information, despite being susceptible to drift caused by both systematic and non-systematic errors. The calibrations of the robot's systematic parameters were conducted using nonlinear optimization and Levenberg-Marquardt methods as Newton-Gauss and gradient-based model fitting methods, while Genetic algorithm and Particle swarm were used as stochastic based methods for kinematic parameter calibration. Performance and results of the calibration methods and pose estimators were validated in detail with experiments on the experimental mobile wall climbing robot.

Autonomous surface vessels (ASVs) are increasingly vital for marine science, offering robust platforms for underwater mapping and inspection. Accurate state estimation, particularly of vehicle pose, is paramount for precise seafloor mapping, as even small surface deviations can have significant consequences when sensing the seafloor below. To address this challenge, we propose an Invariant Extended Kalman Filter (InEKF) framework designed to integrate partial orientation measurements. While conventional estimation often relies on relative position measurements to fixed landmarks, open ocean ASVs primarily observe a receding horizon. We leverage forward-facing monocular cameras to estimate roll and pitch with respect to this horizon, which provides yaw-ambiguous partial orientation information. To effectively utilize these measurements within the InEKF, we introduce a novel framework for incorporating such partial orientation data. This approach contrasts with traditional InEKF implementations that assume full orientation measurements and is particularly relevant for planar vehicle motion constrained to a "seafaring plane." This paper details the developed InEKF framework; its integration with horizon-based roll/pitch observations and dual-antenna GPS heading measurements for ASV state estimation; and provides a comparative analysis against the InEKF using full orientation and a Multiplicative EKF (MEKF). Our results demonstrate the efficacy and robustness of the proposed partial orientation measurements for accurate ASV state estimation in open ocean environments.

To overcome the limitation of existing indoor odometry technologies which often cannot simultaneously meet requirements for accuracy cost-effectiveness, and robustness-this paper proposes a novel magnetometer array-aided inertial odometry approach, MSCEKF-MIO (Multi-State Constraint Extended Kalman Filter-based Magnetic-Inertial Odometry). We construct a magnetic field model by fitting measurements from the magnetometer array and then use temporal variations in this model-extracted from continuous observations-to estimate the carrier's absolute velocity. Furthermore, we implement the MSCEKF framework to fuse observed magnetic field variations with position and attitude estimates from inertial navigation system (INS) integration, thereby enabling autonomous, high-precision indoor relative positioning. Experimental results demonstrate that the proposed algorithm achieves superior velocity estimation accuracy and horizontal positioning precision relative to state-of-the-art magnetic array-aided INS algorithms (MAINS). On datasets with trajectory lengths of 150-250m, the proposed method yields an average horizontal position RMSE of approximately 2.5m. In areas with distinctive magnetic features, the magneto-inertial odometry achieves a velocity estimation accuracy of 0.07m/s. Consequently, the proposed method offers a novel positioning solution characterized by low power consumption, cost-effectiveness, and high reliability in complex indoor environments.

--As robotics progresses toward general manipulation, dexterous hands are becoming increasingly critical. However, proprioception in dexterous hands remains a bottleneck due to limitations in volume and generality. In this work, we present HandCept, a novel visual-inertial proprioception framework designed to overcome the challenges of traditional joint angle estimation methods. HandCept addresses the difficulty of achieving accurate and robust joint angle estimation in dynamic environments where both visual and inertial measurements are prone to noise and drift. T o support sim-to-real transfer, we also open-sourced our high-fidelity rendering pipeline, which is essential for training without real-world ground truth. This work offers a robust, generalizable solution for proprioception in dexterous hands, with significant implications for robotic manipulation and human-robot interaction. Performing these tasks often involves manipulation--either bimanual or in-hand dexterous manipulation--to interact with the environment. The end-effector plays a fundamental role in enabling such robotic manipulation [3]-[5].