Fig. 1. PlantMax planting machine built on Eco Log 574E forwarder, equipped with a MidiFlex disc trencher. Photo Mikael Rönnqvist.

Fig. 2. The PlantMax GNSS tracks shown in gray across test sites A–C, with the Depth-to-Water map in blue and the slope magnitude map in red. The black lines represent the site borders, and the brown contours indicate elevation in meters.

Fig. 3. Overview of the PlantP-module input (a-d) and output (e-f) used to define the net area for regeneration. a) Overview map of the test site with boarders of the planned area to harvest including No-go areas defined by Holmen Skog (red areas). b) Harvester positions when feeling a tree from harvested production (hpr) files are used to define the net regeneration area. (c) Depth-to-Water (DTW) map where a 0.3 m threshold is used to exclude areas predicted to be too wet for mechanical site preparation. (d) Parent material maps from the Swedish Geological Survey are used to exclude peat soils. (e) Net area for regeneration including wet areas f) Net area for regeneration excluding wet areas. Manually suggested passages in wet areas are marked with yellow lines. © Lantmäteriet, © The Swedish Environmental Protection Agency, © Swedish Geological Survey.

Fig. 4. (a) TerraTrail’s node grid of site C, with every third node column sharing a split index, displayed by the color of the node. (b) The Euclidean path segments created by the three different TSP calculations, defined by the split index. (c) The line segments united into one, and a Dubins path has been applied, which considers the turning radius of the vehicle.

Fig. 5. The two distinct Dubins paths generated from the Euclidean Concorde TSP sequence for TerraTrail. The simpler Dubins path, which was calculated first, requires a detour at the upper-left corner to tangent the upper-left node with a straight leftward Dubins heading, as defined by Eq. 1. By calculating the Enhanced Dubins path, altering the Dubins heading of the two upper nodes marked with red arrows, resulted in a 30% shorter path, compared to the simple Dubins path. This resulted in changing the Dubins heading direction of the upper-left node to a downward orientation.

Fig. 6. Results from applying the Pathfinder solution approach to site C. (a) distribution of hexagons in the area, with red hexagons representing no-go zones, (b) initial dominating slope aspects indicated by arrows, (c) solution of mixed sized hexagons with similar slope aspects in the first phase, (d) identified regions with similar aspects in different colors, (e) TSP-solution (f) full route solution with turning and driving direction indicated. Pink indicates turns smaller than the turning radius where reversing is needed.

Fig. 7. The coarse dashed Euclidean connections is exchanged with Dubins paths marked with a fine dotted line, calculated with a heading which is tangent to the driving lines marked with the solid lines.

Fig. 8. GNSS track of PlantMax, along with the two automated path planners, TerraTrail and Pathfinder, on the three test sites of this study. This test scenario is unconstrained and aims to cover the entire site with the shortest path possible.

Fig. 9. GNSS track of PlantMax, along with the two automated path planners, TerraTrail and Pathfinder, on the three test sites of this study. This test scenario is constrained to the Depth-to-Water (DTW) map and aims to cover the entire site with the shortest path possible, while avoiding wet (blue) areas. Note that the Pathfinder route is only using the forcing passages in Fig. 3f, even though it looks like the passages are in various places. These lines are normally dotted to make the operator aware that it is not the exact path visualized.