Journal of Intelligent and Robotic Systems

In recent years 3D models of buildings are used in maintenance and inspection, preservation, and other building related applications. However, the usage of these models is limited because most models are pure representations with no or little associated semantics. In this paper we present a pipeline of techniques used for interior interpretation, object detection, and adding energy related semantics to windows of a 3D thermal model. A sequence of algorithms is presented for building the fundamental semantics of a 3D model. Among other things, these algorithms enable the system to differentiate between objects in a room and objects that are part of the room, e.g. floor, windows. Subsequently, the thermal information is used to construct a stochastic mathematical model– namely Markov Random Field– of the temperature distribution of the detected windows. As a result, the MAP(Maximum a posteriori) framework is used to further label the windows as either open, closed or damaged based upon their temperature distribution. The experimental results showed the robustness of the techniques. Furthermore, a strategy to optimize the free parameters is described, in cases where there is a sample training dataset.

Faults or failures are inevitable to occur and their prompt detection and isolation are essential for the dependability of various systems and for avoiding damages to the system itself, persons and the environment. Therefore, the safety of helicopter platforms have attracted the attention of many researchers in the past two decades. In order to deal with these problems, this paper presents an overview of the recent development and current researches in the field of fault diagnosis, including analytical/model-based, signal processing-based and knowledge-based techniques, and also passive/active fault- tolerant control approaches. Among various helicopters, single-rotor aerial vehicles, i.e. manned helicopters, unmanned helicopters, two and three degree-of-freedom unmanned helicopter experimental platforms, are considered for providing an overall picture of the fault diagnosis and fault-tolerant control approaches based on the review of journal articles in last two decades, conference articles in last several years and some books.

In the last decade, the collision avoidance of Unmanned Aerial Vehicles (UAVs) has become increasingly important for the safe operation of UAVs. In this article, an effective conflict detection and alerting principle is firstly proposed based on mixed collision cone and alerting criterion for the collision avoidance of UAV. Second, a reactive collision avoidance and trajectory recovery strategy (RCATRS) is presented based on the model of relative kinematics. In this strategy, by acting an acceleration vector with different magnitude and direction on UAV, a horizontal collision avoidance maneuver is realized in situations of different relative position and velocity vector. When UAV has bypassed the obstacle, the horizontal trajectory recovery maneuver is initiated to make UAV to return to original flight paths. Thus, different recovery trajectories corresponding with different relative velocity vector are planned. Finally, the safety controller is designed to apply RCATRS to autonomous quadrotor. Since a few of computation is off-line, RCATRS is simple in implement and can satisfy the request of running in real time. The results of simulation show the validity of the proposed RCATRS.

This paper presents a cooperative navigation technique which exploits relative vision-based sensing and carrier-phase differential GPS (CDGPS) among antennas embarked on different flying platforms, to provide accurate UAV attitude estimates in real time or in post-processing phase. It is assumed that all UAVs are under nominal GPS coverage. The logical architecture and the main algorithmic steps are highlighted, and the adopted CDGPS processing strategy is described. The experimental setup used to evaluate the proposed approach comprises two multi-rotors and two ground antennas, one of which is used as a benchmark for attitude accuracy estimation. Results from flight tests are presented in which the attitude solution obtained by integrating CDGPS and vision (CDGPS/Vision) measurements within and Extended Kalman Filter is compared with estimates provided by the onboard navigation system and with the results of a formerly developed code-based differential GPS (DGPS/Vision) approach. Benchmark-based analyses confirm that CDGPS/Vision approach outperforms both onboard navigation system and DGPS/Vision approach.

This paper describes an integration strategy based upon a modular architecture which is meant to improve access to assistive technical devices in the rehabilitation field. This system concept is now known as M3S: Multiple Master Multiple Slave. With M3S, it is possible to connect input devices (like joysticks and keyboards) to end-effectors (like wheelchairs, robots and infra-red remote controllers) to form an integral aid which offers disabled people better opportunities to function as independently as possible. Since M3S is based upon a modular architecture, it allows users (disabled people, attendants, therapists) to compile a specific package of any combination of technical aids to a complete integral system, while still permitting them to extend or modify the system later on. Furthermore the system can be used right-away without any special adaptations using the M3S plug-and-play capabilities. The power of such an integrated system have been shown in several user evaluations in various countries around Europe. The M3S specification is an open standard available for free, M3S has also been proposed to the ISO for formal standardization. For the development of M3S devices a complete set of software tools is available at no cost, hardware starter kits are available for a small fee. Information about M3S can be acquired from the M3S web server (http://www.tno.nl/m3s) or directly from the M3S Dissemination office.

In this paper, an imitation learning approach of vision guided reaching skill is proposed for robotic precision manipulation, which enables the robot to adapt its end-effector’s nonlinear motion with the awareness of collision-avoidance. The reaching skill model firstly uses the raw images of objects as inputs, and generates the incremental motion command to guide the lower-level vision-based controller. The needle’s tip is detected in image space and the obstacle region is extracted by image segmentation. A neighborhood-sampling method is designed for needle component collision perception, which includes a neural networks based attention module. The neural network based policy module infers the desired motion in the image space according to the neighborhood-sampling result, goal and current positions of the needle’s tip. A refinement module is developed to further improve the performance of the policy module. In three dimensional (3D) manipulation tasks, typically two cameras are used for image-based vision control. Therefore, considering the epipolar constraint, the relative movements in two cameras’ views are refined by optimization. Experimental are conducted to validate the effectiveness of the proposed methods.

In this paper, we develop a two-dimensional articulated body tracking algorithm based on the particle filtering method using partitioned sampling and model constraints. Particle filtering has been proven to be an effective approach in the object tracking field, especially when dealing with single-object tracking. However, when applying it to human body tracking, we have to face a “particle-explosion” problem. We then introduce partitioned sampling, applied to a new articulated human body model, to solve this problem. Furthermore, we develop a propagating method originated from belief propagation (BP), which enables a set of particles to carry several constraints. The proposed algorithm is then applied to tracking articulated body motion in several testing scenarios. The experimental results indicate that the proposed algorithm is effective and reliable for 2D articulated pose tracking.

In this paper, we present an intelligent and flexible framework for autonomous pick-and-place tasks in previously unknown scenarios. It includes modules for object recognition, environment modeling, motion planning and collision avoidance, as well as sophisticated error handling and a task supervisor. The framework combines state-of-the-art algorithms and was validated during the first phase of the European Robotics Challenge in which it obtained first place in a field of 39 international contestants. We discuss our results and the potential application of our framework to real industrial tasks. Furthermore, we validate our approach with an application on a real harvesting manipulator. To inspire other teams participating in the challenge and as a tool for new researchers in the field, we release it as open source.

The aircraft collision avoidance problem can be formulated using a decision-theoretic planning framework where the optimal behavior requires balancing the competing objectives of avoiding collision and adhering to a flight plan. Due to noise in the sensor measurements and the stochasticity of intruder state trajectories, a natural representation of the problem is as a partially-observable Markov decision process (POMDP), where the underlying state of the system is Markovian and the observations depend probabilistically on the state. Many algorithms for finding approximate solutions to POMDPs exist in the literature, but they typically require discretization of the state and observation spaces. This paper investigates the introduction of a sample-based representation of state uncertainty to an existing algorithm called Real-Time Belief Space Search (RTBSS), which leverages branch-and-bound pruning to make searching the belief space for the optimal action more efficient. The resulting algorithm, called Monte Carlo Real-Time Belief Space Search (MC-RTBSS), is demonstrated on encounter scenarios in simulation using a beacon-based surveillance system and a probabilistic intruder model derived from recorded radar data.