This study presents a metaheuristic approach to coverage path planning for ground-based forest operations, focusing on minimizing path lengths for forest vehicles while considering terrain characteristics and vehicle parameters. Forest vehicles can usually tolerate higher pitch than roll angles, which makes them vulnerable to rollover. To mitigate that, this method utilizes a genetic algorithm to optimize the sequence of nodes, which are scattered over the site with equal spacing. The coverage path planner then calculates the Dubins path distance between every node in the fitness function, together with penalties for exceeding pitch, roll and soil moisture constraints for the vehicle. This ensures that the path planner tries to make the most traversable path as possible, while trying to minimize the driving distance. Two synthetic test sites resembling primitive challenging terrains, and one real site were utilized to theoretically evaluate the proposed method. The results show that aligning the node patterns with the critical slope headings, instead of having a straight pattern, had little effect on the path length. However, square grids can yield shorter paths across multiple runs, while triangular grids ensure consistent results in single runs. A two-hectare site took 43 minutes to calculate on average. This suggests that further development of the path planner could lead to significant improvements, enabling the management of sites larger than a few hundred nodes. However, the calculation time is justified for the reduced path length during deployment. The study presents a methodology that supports manual operators and establishes foundations for full autonomy.

Sustainable forestry requires efficient regeneration methods to ensure that new forests are established quickly. In Sweden, 99% of the planting is manual, but finding labor for this arduous work is difficult. An autonomous scarifying and planting machine with high precision, low environmental impact, and a good work environment would meet the needs of the forest industry. For two years, a collaborative group of researchers, manufacturers, and users (forest companies) has worked together on developing and testing a new concept for autonomous forest regeneration (Autoplant). The concept comprises several subsystems, i.e., regeneration and route planning, autonomous driving (path planning), new technology for forest regeneration with minimal environmental impact, automatic plant management, crane motion planning, detection of planting spots, and follow-up. The subsystems were tested separately and integrated together during a field test at a clearcut. The concept shows great potential, especially from an environmental perspective, with significantly reduced soil disturbances, from approximately 50% (the area proportion of the area disturbed by disc trenching) to less than 3%. The Autoplant project highlights the challenges and opportunities related to future development, e.g., the relation between machine cost and operating speed, sensor robustness in response to vibrations and weather, and precision in detecting the size and type of obstacles during autonomous driving and planting.

An integral part of autonomous forestry is the ability of the vehicles, e.g., forwarders and harvesters, to perceive their environment. At Luleå University of Technology, object detectors have previously been developed, allowing forestry vehicles to detect and position important objects in forestry, such as tree stumps, stones, and logs. These detectors have been developed by training on physical manually annotated data, which is both time-consuming and costly. Training on virtual data allows for significant time- and cost reductions. Since the ground truth in the virtual model is known, the training data can be auto-annotated, allowing for the creation of larger training datasets, at a lower cost. In this work, a virtual environment in Unity is used in co-simulation with a real-time digital twin of a physical forestry vehicle, to generate auto-annotated training data, as captured by an onboard stereo camera. A detailed emulation of the stereo camera is used to achieve realistic results. First, a log detector trained on physical manually annotated data, is evaluated on virtually created data. It is shown that the log detector trained on physical data can detect logs in the virtual environment. Second, new detectors are trained, using different shares of physical and virtual data. It is shown that a detector trained using only virtual data, can learn to detect logs in the physical world. Moreover, virtual pre-training is shown to improve the performance of physically trained and tested detectors, both at low availability of physical training data, and in terms of domain generalization. Furthermore, the real-time capable virtual models also enable future machine learning tasks utilizing different levels of Hardware-in-the-Loop.

This work presents outdoor test results of autonomous navigation system based on two different reference points of PurePursuit algorithm for a 10-ton articulated research platform. PurePursit is commonly used simplified tracking algorithm based on geometric calculation of desired steering angle by pursuing certain number of points/distance ahead on given path from a fixed reference point of vehicle. Choice of fixed reference point effects the tracking accuracy particularly for articulated/center-steered vehicles. In this experimental work, PurePursuit algorithm with a virtual reference point(PPV) and commonly used front-axle reference point(PPF) is evaluated for heavy duty articulated vehicles in outdoor experiments. Experimental data shows that choice of reference point in PurePursuit algorithm for articulated vehicle has impact on tracking accuracy in terms of crosstrack errors and heading errors. Navigation tests were performed on a flat asphalt surface for paths of sharp complexities i-e a path with continuous curvature (circular path) and a path with sharp turns (zigzag path) with different initial conditions i-e initial position of vehicle. In general, it can be concluded that PurePursuit algorithm with front reference point (PPF) produced fewer crosstrack errors while PurePursuit algorithm with virtual midpoint reference (PPV) produced fewer heading errors.

Accurate autonomous navigation for off-terrain utility vehicles with no human intervention is an essential requirement to achieve full automation. Within several applications, though higher autonomous navigational accuracy (almost ±2.5 cm) has been achieved in some commercially available vehicles yet requirements for human intervention is still very much required in several situations. This study investigates autonomous navigational accuracy of PurePursuit algorithm for different reference points for a non-steerable-wheels center-steered articulated vehicle. PurePursuit algorithm is preferable choice for path tracking for its simplicity and for vehicles where high speed is not a requirement. Evaluation of PurePursuit algorithm for utility vehicles of type studied in this paper is somewhat less explored area. We have compared the autonomous navigational accuracy of PurePursuit algorithm for different reference points for a set of different path complexities that also includes vehicles kinematic constraints in simulation environment. Average lateral and heading deviations were calculated for a set of different path complexities, and it was found that in general proposed reference point for PurePursuit algorithm for center-steer articulated vehicle shows better lateral and heading autonomous navigational accuracy than the traditional PurePursuit algorithm.

Modern forestry practices are based on the idea of ‘big is beautiful’. Especially in the regeneration phase, the operations are often excessive in relation to the profit that one can expect to gain in decades to come. Excessive operations also constrain the use of ecosystem services. Lean forestry is a novel philosophy of forestry practise that aims to direct the idea of “big is beautiful” in modern silviculture more into “do cost effectively only what is needed to fulfil the goals”. To succeed Lean forestry requires exact spatial information to be able to carry out forestry measures very precisely only where they are really needed to fulfil goals. This kind of a paradigm shift requires systems with new kinds of abilities to remotely sense the surrounding environment and to make better and faster decisions based on sensed data. Automated unmanned offroad vehicle that is able to sense the environment and to make lean decisions is presented as an example of initiatives that can make forestry more cost-effective and simultaneously improve utilisation of wide range of ecosystem services in forests.

Abstract: At Luleå University of Technology, an offroad vehicle capable of performing autonomous forwarding of full-scale logs, has been developed and built. This vehicle serves as a modular research platform, where different hardware, software, and algorithms can be exchanged, tested, and validated. To further facilitate this research, a real-time multi-physics digital twin of the vehicle has been developed. The digital twin can be executed either in fully virtual or in Hardware-in-the-Loop mode. Virtual and hybrid testing and development are expected to increase the efficiency, safety, and sustainability of these activities. Future envisioned use-cases also include machine learning and generation of auto-annotated detector training data. The digital twin features multibody simulation, hydraulics, contacts and collisions, flexible tires, emulated sensors, and co-simulation capabilities. The test environment is a physical environment that has been recreated digitally, using data from a UAV-mounted camera and lidar, as well as handheld RTK GPS measurements. Therefore, accurate GPS positions can be obtained from the digital environment. The digital twin is equipped with an emulated GPS sensor and can thus follow the same GPS tracks as the physical vehicle. This paper presents the development and verification of the digital twin model. Calibration and testing of the digital twin’s hydraulically actuated articulated joint is detailed, and the results are analyzed. The modeled hydraulics of the articulated joint show realistic dynamics. It is also shown that the level of detail of the wheel motors’ hydraulic circuits, affects the calibration of the hydraulics model of the articulated joint, as they affect the resistance during turning.