Aerial navigation in obstructed environments with embedded nonlinear model predictive control

Abstract

We propose a methodology for autonomous aerial navigation and obstacle avoidance of micro aerial vehicles (MAVs) using non-linear model predictive control (NMPC) and we demonstrate its effectiveness with laboratory experiments. The proposed methodology can accommodate obstacles of arbitrary, potentially non-convex, geometry. The NMPC problem is solved using PANOC: a fast numerical optimization method which is completely matrix-free, is not sensitive to ill conditioning, involves only simple algebraic operations and is suitable for embedded NMPC. A C89 implementation of PANOC solves the NMPC problem at a rate of 20 Hz on board a lab-scale MAV. The MAV performs smooth maneuvers moving around an obstacle. For increased autonomy, we propose a simple method to compensate for the reduction of thrust over time, which comes from the depletion of the MAV's battery, by estimating the thrust constant.

Where to look: a collection of methods forMAV heading correction in underground tunnels

“…Additionally, the potential fields method [33] is implemented to generate velocity references , to avoid collisions to the local surrounds using range measurements R of the 2D lidar placed on top of the MAV. Furthermore, for tracking the desired velocity and altitude references the Non‐linear Model Predictive Control (NMPC) [34] is implemented to generate the corresponding thrust and attitude commands for the low level controller. The low level controller generates the motor commands for the MAV.…”

Non-linear MPC based Navigation for Micro Aerial Vehicles in Constrained Environments

“…In the related literature, there are numerous control schemes for path-planning and obstacle avoidance that have successfully been utilized for MAV's [6], [7], for example reactive navigation schemes as the widely used potential fields [8], [2] or dynamic graph search methods like the modified versions of A such as ADA* [9] or MSA* [10]. Lately, there have been a few approaches towards NMPC schemes for the combined problem of control and path planning as in [11], while the ability has been demonstrated of such advanced and demanding control schemes to operate efficiently in experimental environments [12]. In the specific case of the NMPC schemes, the main problematic issue is the solver time and the main ability to compute in real-time or close to real-time way-points that will satisfy the overall mission demands and timing issues.…”

“…The parameter p includes the initial condition of the state vector X as well as references and obstacle data. The idea is to re-solve (12) with an exponentially increasing penalty parameter until the constraints satisfy a specified tolerance. This can be seen as the cost-minima gradually moving out of the obstacles as c is increased until the path does not enter the obstacles.…”

A Unified NMPC Scheme for MAVs Navigation With 3D Collision Avoidance Under Position Uncertainty

“…Few articles have addressed collision avoidance of MAVs with the NMPC framework. Much research focused on addressing formation problems with a fixed number of agents [19], considered a linear model [20], assumed global pose information of obstacles [21], uncertainties of position estimations are excluded.…”