Pekka Forsman

3-D mapping of natural environments with trees by means of mobile perception

In this paper, a method of generating a three-dimensional (3-D) geometric model for large-scale natural environments with trees is presented. The environment mapping method, which uses range images as measurement data, consists of three main phases. First, geometric feature objects are extracted from each of the range images. Second, the relative coordinate transformations (i.e., registrations) between the sensor viewpoint locations, where the range data are measured, are computed. Third, an integrated map is formed by transforming the submap data into a common frame of reference. Tree trunks visible in the range images are modeled with cylinder segments and utilized as reference features for registration computation. The final integrated 3-D model consists of the cylinder segments representing the visible sections of the tree trunks, as well as of the ground elevation data. The constructed environment map can be utilized as, for example, a virtual task environment for outdoor robotic machines such as new-generation forest working machines or service robots.

Simultaneous Localization and Mapping for Forest Harvesters

A real-time SLAM (simultaneous localization and mapping) approach to harvester localization and tree map generation in forest environments is presented in this paper. The method combines 2D laser localization and mapping with GPS information to form global tree maps. Building an incremental map while also using it for localization is the only way a mobile robot can navigate in large outdoor environments. Until recently SLAM has only been confined to small-scale, mostly indoor, environments. We try to addresses the issues of scale for practical implementations of SLAM in extensive outdoor environments. Presented algorithms are tested in real outdoor environments using an all-terrain vehicle equipped with the navigation sensors and a DGPS receiver.

Tree Measurement and Simultaneous Localization and Mapping System for Forest Harvesters

For the last decades, measurement and automation systems in Nordic cut-to-length forestry machines have evolved gradually. These heavy duty machines are lighter, faster and more accurate than ever before but the basic technologies and operation have remained the same. In many respects, their current automation systems have reached their limits. The Forestrix project studies how advances in mobile robotics could be applied in the field of forestry machine automation. Machine vision systems and scanning laser range finders have established themselves as standard equipment in mobile robotics. With the new sensor and computing technologies it is possible to get information about the surrounding forest, such as tree diameters, positions and stand density. This information can be used on-line in operator’s decision support system, or off-line in a forest asset management system. This paper describes the prototype measurement platform and the software algorithms developed in the Forestrix project. Results from tests with an all terrain vehicle are also presented.

Best-first branch and bound search method for map based localization

To know the pose of the robot is one of the central requirements in many applications. A localization algorithm should be robust, it should give the estimate of unreliability and it should be able to recover from errors. The above requirements are often trade offs with the computational complexity. This paper presents a global localization algorithm that matches a local point map (acquired e.g. with a laser range finder) to a global map. The algorithm makes a discrete search using a best-first branch and bound method to efficiently compute the globally optimal pose estimate. Moreover, the degree of ambiguity of the pose estimate can be determined as the algorithm yields all potential pose solution candidates within the search space. Experimental results are given to show the behavior and performance analysis of the algorithm.

Dynamic modeling and obstacle-crossing capability of flexible pendulum-driven ball-shaped robots

Abstract Ball-shaped robots present a novel and widely studied approach for mobile robotics. Despite the essential benefit of the ball-robot that it cannot flip over or fall down, the robot’s physical construction often severely limits the ball mobility in uneven terrain. The customarily applied quasi-static motion model makes the anticipated theoretical robot mobility even worse, because it completely ignores ball dynamics and therefore seriously under-estimates the robot’s obstacle-crossing capability. The energy-based model, sometimes applied instead of the quasi-static model, over-estimates ball mobility and becomes inconvenient when an active driving motor is added to the system. This paper introduces a new extended dynamic model for flexible pendulum-driven ball-shaped robots, as well as a simulation-based method to predict the robot’s step-crossing capability. The extended dynamic model allows rolling, bouncing and slipping of the robot, and it includes a simplified contact model for the ball-obstacle-interaction. The simulation results have been compared to experimental results obtained with a physical robot. The comparison shows that the new dynamic model and contact model outperform the traditionally applied quasi-static and energy-based models. The new dynamic model may be applied in mobility analysis of ball-robot designs, for path planning, as well as for control algorithm development.

Feature based registration of range images for mapping of natural outdoor environments

Challenges related to viewpoint registration in rough forest terrain can be quite different compared to those faced in structured environments. Maneuvring the sensor between measurement positions introduces large error into the a priori estimates of the registration coordinates. As a consequence, locally optimal registration methods may not work properly. Moreover, due to the clutter, the scene contents can change substantially even due to a relative small displacement of the sensor. Often, the sensor has to be moved to the other side of the target object, such as a group of trees, to get a good coverage of its geometry. In both cases, overlap between the two 3D data sets will be minimal ruling out conventional registration methods. In this paper, a feature-based method for registering 3D range scans for mapping natural outdoor environments is proposed. The method utilizes cylindrical, rotation symmetric features extracted from the 3D measurement data for viewpoint registration. The method is tested on real range images.

Unified Representation of Decoupled Dynamic Models for Pendulum-Driven Ball-Shaped Robots

Dynamic models describing the ball-robot motion form the basis for developments in ball-robot mechanics and motion control systems. For this paper, we have conducted a literature review of decoupled forwardmotion models for pendulum-driven ball-shaped robots. The existing models in the literature apply several different conventions in system definition and parameter notation. Even if describing the same mechanical system, the diversity in conventions leads into dynamic models with different forms. As a result, it is difficult to compare, reproduce and apply the models available in the literature. Based on the literature review, we reformulate all common variations of decoupled dynamic forward-motion models using a unified notation and formulation. We have verified all reformulated models through simulations, and present the simulation results for a selected model. In addition, we demonstrate the different system behavior resulting from different ways to apply the pendulum reaction torque, a variation that can be found in the literature. For anyone working with the ball-robots, the unified compilation of the reformulated dynamic models provides an easy access to the models, as well as to the related work.

Dynamic Obstacle Overcoming Capability of Pendulum-driven Ball-Shaped Robots

This paper discusses dynamic step-crossing capability of pendulum-driven ball-shaped robots. We introduce an extended dynamic model that allows modeling of ballrobot rolling, bouncing and slipping. Based on the new model, our simulations predict the maximum over-passable step-height for the robot. The simulation results agree well with the result from a parallel simulation in Adamssoftware as well as with practical experiments. The new dynamic model can be applied for mobility analysis of robot-ball designs as well as for path planning.

Using Signs for Configuring Work Tasks of Service Robots

Purpose – This paper seeks to describe how signs can be used as a part of the work task scenarios with service robot.Design/methodology/approach – The signs are used to control a work task of a mobile service robot. Tests are done in real outdoor and non‐structured environments with the WorkPartner robot. In the test scenario the signs are used to define the boundaries of a working area. Two different kinds of signs are tested: passive and active.Findings – The signs can be used as a simple way to define the parameters of a work task. One of the advantages is usability. For example, the signs can just be carried to the working area and no other actions are needed.Research limitations/implications – The tests focused only on bounding the working area with signs. Using the signs for pointing a direction, marking a route, or defining a location of a single target will be tested in the future.Practical implications – Signs are an easy way to configure work tasks of service robots. The concept can be applied t...

Gyroscopic Precession In Motion Modelling Of Ball-Shaped Robots

This study discusses kinematic and dynamic precession models for a rolling ball with a finite contact area and a point contact respectively. In literature, both conventions have been applied. In this paper, we discuss in detail the kinematic and dynamic models to describe the ball precession and the radius of a circular rolling path. The kinematic model can be used if the contact area and friction coefficient are sufficient to prevent slippage. The dynamic precession model has significance in multi-body simulation environments handling rolling balls with ideal point contacts. We have applied both the kinematic and dynamic precession model to evaluate the no-slip condition of the existing GimBall-robot. According to the result, the necessity of an external precession torque may cause slipping at lower velocities than expected if ignoring this torque.