Autonomous aerial navigation using monocular visual‐inertial fusion

Abstract

Autonomous micro aerial vehicles (MAVs) have cost and mobility benefits, making them ideal robotic platforms for applications including aerial photography, surveillance, and search and rescue. As the platform scales down, MAVs become more capable of operating in confined environments, but it also introduces significant size and payload constraints. A monocular visual-inertial navigation system (VINS), consisting only of an inertial measurement unit (IMU) and a camera, becomes the most suitable sensor suite in this case, thanks to its light weight and small footprint.In fact, it is the minimum sensor suite allowing autonomous flight with sufficient environmental awareness. In this paper, we show that it is possible to achieve reliable online autonomous navigation using monocular VINS. Our system is built on a customized quadrotor testbed equipped with a fisheye camera, a low-cost IMU, and heterogeneous onboard computing resources. The backbone of our system is a highly accurate optimization-based monocular visual-inertial state estimator with online initialization and self-extrinsic calibration. An onboard GPU-based monocular dense mapping module that conditions on the estimated pose provides wide-angle situational awareness. Finally, an online trajectory planner that operates directly on the incrementally built threedimensional map guarantees safe navigation through cluttered environments. Extensive experimental results are provided to validate individual system modules as well as the overall performance in both indoor and outdoor environments.

A Survey on Aerial Swarm Robotics

“…Such representations can be sparse [184], semidense [185], or fully dense [186]. While dense representations can be directly used for autonomous navigation [186] or geographical reference, sparse representations are often only used for state estimation [187] or collaborative control of robotic agents. Due to the fact that the environment is often only partially known or even totally unknown, mapping tasks are often coupled with localization (pose estimation) issues, which turn them into the classic simultaneous localization and mapping (SLAM) problem.…”

Autonomous Flight

“…Recent work has shown that obstacle detection for reactive collision avoidance is possible at high speeds on a CPU (68) running onboard a fixed-wing aircraft traveling at up to 14 m/s. However, maintaining a complete map for longer-range planning is computationally expensive, and as a result, most mapping algorithms have been executed using either offboard resources (50) or onboard specialized hardware, such as GPUs (53) or field-programmable gate arrays (69). These mapping algorithms can be combined with trajectory-generation methods to enable autonomous navigation through cluttered environments (49,53,68).…”

“…Many works have successfully demonstrated autonomous navigation in unknown workspaces using completely onboard mapping, estimation, planning, and control pipelines (53,66,85,86,90) (49,85), which uses a stereo camera and IMU for estimation and a laser sensor for mapping, can reach speeds of 7 m/s. Figure 5 illustrates identified safe trajectories from a single start position to multiple goals throughout a space.…”

Monocular Visual-Inertial State Estimation for Micro Aerial Vehicles

“…The dimensions of a small platform also limit its ability to carry stereo or multicamera systems due to insufficient baseline length. As the platform becomes smaller, a monocular visualinertial navigation system (VINS), consisting of only a low-cost inertial measurement unit (IMU) and a camera, becomes the only viable sensor suite allowing autonomous flights with sufficient environmental awareness [4].…”

“…Recent research on VINS for MAVs have yielded a number of significant results [1][2][3][4]11]. Most of these existing approach can be classified into filter-based and optimization-based systems.…”