Projects sharing features with BEVFormer

OM1 is a multimodal AI agent runtime and orchestration framework designed to connect large language models to physical robot hardware and sensors. It provides an execution environment that processes audio, video, and sensor data to drive autonomous decisions and actions in real-world settings. The system integrates a robotics SLAM and navigation stack with a hardware abstraction layer, allowing high-level AI commands to be translated into low-level motor and actuator instructions. It distinguishes itself by incorporating blockchain-based governance to enforce immutable operational rules and p

ARKit-CoreLocation is a location integration framework designed to synchronize augmented reality scenes with real-world GPS coordinates. It provides a geospatial positioning engine and anchor system that links 3D scene objects to specific physical latitudes and longitudes. The framework distinguishes itself through sensor-fusion coordinate alignment, combining high-frequency motion data with GPS updates to maintain stable coordinates and refine user positioning. It includes specialized tools for compass calibration and scene orientation correction to align the 3D environment with True North.

Autoware is a modular autonomous driving stack and open-source platform for advanced driver assistance systems. It functions as an integrated operating environment that manages the full pipeline from sensor data processing to vehicle actuation, utilizing the ROS 2 robotics framework for distributed communication and hardware abstraction. The system provides a comprehensive software architecture to enable autonomous driving across various vehicle platforms. It coordinates perception, planning, and control systems to operate vehicles without human intervention. The platform covers several core

CenterNet is a center point object detection framework and real-time computer vision pipeline. It identifies objects and poses by predicting center points instead of using anchor boxes. The system functions as a 3D bounding box estimator, a human pose estimation model, and a tool for real-time object detection. It treats the placement of joints and object locations as center point detection problems to locate entities in images and three-dimensional space. Capabilities cover 3D object detection, human keypoint estimation, and live video analysis. The pipeline uses a single-stage feedforward

Cartographer is a software library and spatial localization engine for simultaneous localization and mapping. It provides a framework for calculating the precise position and orientation of a device while concurrently generating real-time 2D and 3D representations of its environment using lidar-based data. The system implements a real-time mapping approach that uses live sensor streams to track device heading and position. It utilizes a submap-based mapping strategy to divide environments into local maps that are aligned into a global map. The project covers a range of SLAM capabilities, inc

Gyroflow is a gyroscope video stabilization software and IMU telemetry processor designed to remove camera shake from video files. It functions as a hardware-accelerated video renderer and lens calibration tool, utilizing embedded or external gyroscope and accelerometer data to perform pixel-level stabilization. The system is distinguished by its ability to integrate with professional non-linear video editing software via plugins, allowing stabilization to be applied directly to timelines without transcoding original footage. It supports diverse telemetry ingestion from camera brands, flight

FAST-LIVO2 is a LiDAR-inertial odometry framework and factor-graph SLAM implementation designed for real-time robot localization and 3D mapping. It functions as a multi-sensor fusion pipeline and state estimator that integrates LiDAR, inertial, and camera inputs to track a robot's position and orientation. The system employs a tightly-coupled sensor fusion approach to maintain stable navigation, particularly in degraded environments. It utilizes a voxel-based 3D mapping tool to organize point clouds into volumetric grids, which optimizes memory usage and search speed during spatial reconstruc

Grounded-Segment-Anything is a suite of specialized tools for multimodal visual analysis, text-based segmentation, and generative image editing. It integrates text-to-bounding-box detection and high-precision image segmentation masks to function as a text-based image segmenter and an automated visual labeling tool. The project enables text-driven image editing by identifying objects through natural language to perform inpainting and element replacement. It further extends visual analysis into three dimensions, allowing for 3D human reconstruction and the generation of 3D bounding boxes from t

PCDet is a LiDAR 3D object detection toolbox and point cloud processing library built on the PyTorch deep learning framework. It provides a system for identifying and locating three-dimensional objects within point cloud data. The project utilizes a data-model separation pattern to decouple dataset loading logic from the core detection pipeline. It features a multi-sensor fusion pipeline that combines data from multiple sensors into a shared spatial view and a distributed GPU training system to scale workloads across multiple graphics processors. The toolkit covers several capability areas,

PX4-Autopilot is a professional-grade flight control software stack designed for autonomous unmanned vehicles, including multicopters, fixed-wing aircraft, and vertical takeoff and landing platforms. It operates as a modular, real-time framework that decouples flight control logic from hardware drivers through a publish-subscribe middleware architecture. The system utilizes a deterministic microkernel runtime to execute time-critical flight control loops and sensor fusion tasks, ensuring stable navigation and vehicle operation. The platform distinguishes itself through a parameter-driven conf

jetson-inference is a set of libraries and tools for executing optimized deep learning models on embedded GPU hardware. Its primary purpose is to enable real-time computer vision and AI inference at the edge with low latency and high throughput. The project distinguishes itself through high-performance streaming analytics and the ability to execute concurrent AI pipelines on auto-grade silicon. It provides specialized support for multi-sensor stream processing, utilizing zero-copy data transport to load camera frames directly into GPU memory. The codebase covers a broad surface of capabiliti

Openpilot is an open-source driver assistance system that integrates with vehicle control units to provide automated steering, acceleration, and braking. It functions as an automotive robotics middleware, utilizing a specialized runtime environment to process sensor data and execute real-time control commands that manage vehicle dynamics. The platform distinguishes itself through a hardware-agnostic interface that translates standardized driving commands into the proprietary protocols required by a wide range of vehicle makes and models. It employs neural-network-based path planning to predic

Unofficial implemention of lanenet model for real time lane detection

Autoware Universe serves as a foundational pillar within the Autoware ecosystem, playing a critical role in enhancing the core functionalities of autonomous driving technologies. This repository is a pivotal element of the Autoware Core/Universe concept, managing a wide array of packages that…

Udacity Self-Driving Car Engineer Nanodegree projects.

A Kitti Road Segmentation model implemented in tensorflow.

Side-Menu.Android is a reusable UI component for Android applications that provides a slide-out navigation drawer. It is designed to help developers organize application sections and user options into a structured, hidden panel that maintains a clean interface for the primary content area. The component distinguishes itself through its visual presentation, which follows Material Design guidelines to ensure a consistent and intuitive user experience. It features a data-driven menu hierarchy that allows for logical grouping of navigation items, and it incorporates fluid circular reveal animatio

Detectron2 is a PyTorch computer vision framework and visual recognition platform designed for training and deploying models for object detection, image segmentation, and visual recognition. It provides a research-oriented environment for training complex vision models with multi-GPU acceleration. The project includes a specialized object detection library for identifying and locating multiple objects via bounding boxes, as well as an image segmentation toolkit for creating pixel-level masks through instance, semantic, and panoptic segmentation. Additionally, it features a human pose estimati

This project is a robotics software package designed for simultaneous localization and mapping, providing a framework for visual-inertial odometry and environmental mapping. It functions as a middleware-integrated library that enables autonomous mobile robots to estimate their position and orientation by processing sensor data within modular software systems. The library distinguishes itself by utilizing hardware-accelerated processing to perform feature tracking and odometry calculations on dedicated graphics hardware. It maintains spatial accuracy through graph-based optimization and statis

pygta5 is a Python game automation library designed to control actions and simulate player input within Grand Theft Auto 5. It serves as a framework for collecting game data, processing visual frames via neural networks, and automating gameplay through deep learning. The project implements a convolutional neural network controller to make real-time driving and movement predictions based on visual game frames. It utilizes regression models and deep learning to execute autonomous actions, allowing for the creation of autonomous agents that can control characters or vehicles. The system include

Dummy-Robot is a software platform designed for the real-time teleoperation of desktop robotic arms. It functions as a human-robot interaction interface, translating gestural input and sensor data into fluid mechanical motion to enable remote physical task execution. The system utilizes a specialized control framework to map human hand movements directly to robotic actuators. By integrating sensor fusion and kinematic mapping, the software maintains low-latency synchronization between an operator's spatial inputs and the physical assembly, allowing for precise manipulation through a dedicated

Espectre is an edge machine learning framework and motion detection platform that uses Wi-Fi Channel State Information to identify human presence and movement. It functions as a sensing toolkit for ESP32 microcontrollers, enabling the detection of motion through walls without the use of cameras or wearables. The project distinguishes itself by executing compact neural network classifiers and mathematical detection algorithms directly on the microcontroller. It utilizes a MicroPython runtime to allow for the prototyping and deployment of sensing logic and wireless signal processing algorithms

NuttX is a POSIX-compliant real-time operating system designed for microcontrollers ranging from 8-bit to 64-bit architectures. It provides a deterministic execution environment with a real-time task scheduler and a POSIX embedded kernel to ensure portable code execution across diverse hardware targets. The project distinguishes itself through a comprehensive hardware abstraction layer that provides standardized drivers for I2C, SPI, CAN, and USB across various semiconductor chipsets. It also features an embedded networking stack supporting TCP, UDP, IPv4, and IPv6, alongside industrial proto