Contributions published at Artificial Intelligence Lab

null

null

null

null

null

Self-assembly is a phenomenon broadly observed in nature where a vast number of various molecules spontaneously synthesize complex structures. In this article, prompted by the need for the realization of highly autonomous self-assembly systems that employ magnetism as a driving force, we discuss fundamental issues associated with magnetically driven self-assembly systems. We first introduce some examples from our case studies, in which the models all subscribe to a distributed approach, and thus lack central control. Then we categorize them by their type of magnetic attachment. The issues discussed include several fundamental properties, such as the effect of morphology, stochasticity, the difference between two-dimensional models and three-dimensional models, emergence, allostericity, and parallelism. The conclusions obtained support our stance that the appropriate morphology lightens the control cost for the assembly, providing primal but engaging instances of magnetic self-assembly systems that warrant further study.

null

Despite many efforts, balance control of humanoid robots in the presence of unforeseen external or internal forces has remained an unsolved problem. The difficulty of this problem is a consequence of the high dimensionality of the action space of a humanoid robot, due to its large number of degrees of freedom (joints), and of non-linearities in its kinematic chains. Biped biological organisms face similar difficulties, but have nevertheless solved this problem. Experimental data reveal that many biological organisms reduce the high dimensionality of their action space by generating movements through linear superposition of a rather small number of stereotypical combinations of simultaneous movements of many joints, to which we refer as kinematic synergies in this paper. We show that by constructing two suitable non-linear kinematic synergies for the lower part of the body of a humanoid robot, balance control can in fact be reduced to a linear control problem, at least in the case of relatively slow movements. We demonstrate for a variety of tasks that the humanoid robot HOAP-2 acquires through this approach the capability to balance dynamically against unforeseen disturbances that may arise from external forces or from manipulating unknown loads.

The principle introduced in this thesis enables a robot to learn reflex-like behaviours automatically. This is done by performing twitches with a tendon-driven pendulum robot equipped with several sensor modalities (force, length- and tactile sensors) and leads to the emergence of four behaviours (equivalent to myotatic reflex, reverse myotatic reflex, reciprocal inhibition reflex and withdrawal reflex). The reflexes emerge from the structure of the robot with the help of a correlation-based learning scheme. The thesis describes the four reflexes, the robot including its sensors, the learning procedure and the production of the reflexes.

A main drawback in today's upper limb prostheses is the lack of sensory feedback, requiring patients to operate the prosthesis under sight control. A stimulation device, using transcutaneous electrical nerve stimulation has been previously developed at the University of Zurich. The goal was to embed this stimulator in a myoelectric prosthetic control environment for further investigation on the benefit this kind of feedback is able to provide for prosthetic operation. A modular framework was developed, integrating EMG sensors, a robotic hand and the stimulator. The prosthetic hand's force and position information was output onto the user's lower back through one channel each. Experiments showed that the stimulator was able to return perceivable information about the hand's status.

Crowdsourcing is a phenomenon that changes the way organizations use the Internet to collect ideas, solve complex cognitive problems, and build high-quality data and knowledge repositories (e.g. Wikipedia) or even software (e.g., Linux) by self-organizing agents around data and knowledge. Many recent studies have highlighted the factors and the small sets of parameters that play a role when a large crowd interacts with an organization. However, no comprehensive simulation has yet been developed to incorporate all these parameters and potentially generate predictive power. This thesis, based on a multi-agent system, describes the development of a simulator named “CrowdSim”, for human crowds performing collective problem solving in a Crowdsourcing scenario. The simulator allows running sensitivity analyses of multiple parameters as well as simulating intractable interactions of complex networks of irrational agents. In addition, the modular and extensible way the simulator is built enables the user to increase the accuracy and predictive power when scientists gain new empirical insights.

Everyday manipulation of object involves a complex tactile sense system that is able to acquire rich information about objects and influence grasp control. Prosthetic hand wearers, that lost their tactile sense associated to the missing hand, are deprived of this information, thus encountering difficulties in efficiently manipulating objects in their daily activity. The restoration of the lost tactile sense for amputees wearing a prosthetic hand poses various challenges related to the design of interfaces that allow an efficient and rich flow of information between the environment and the prosthetic hand, and between the prosthetic hand and the user. Thus, main endeavors in haptic interfaces focus on developing tactile sensors or haptic devices in order to relay environmental information. The focus of this thesis is to inquire on stable and efficient grasp during object manipulation. To this aim, in this thesis we took the following investigation courses: 1. We developed an artificial ridged skin and tested its potential to extract enriched information (force, slippage detection, slippage speed) from the interaction with a slipping object using one force sensor. The approach aims to offer high information bandwidth to cost ratio, as well as to reduce the complexity (energy, weight) of prosthetic hands. The evaluation of the artificial ridged skin was performed under theoretical and practical conditions. 2. We investigated tactile parameters (force, slippage speed, force and slippage speed) that can lead to optimal grasp during object slippage. The evaluation was performed in a psychophysical experiment with human participants in a combined real and virtual environment. The results show that the artificial ridged skin is a promising tactile sensor for extracting information related to slippage. Furthermore, the results suggest that force and slippage speed feedback may enable an optimal reaction time and grip for overcoming slippage. The study supports a better understanding of the requirements for stable grasp with prostheses. Additionally, the performance of the artificial ridged skin comes to propose the means of reaching a stable grasp.

It has been a quite while since people realized that self-assembly technique may be a strong method to manufacture 3D micro products. In this contribution, we investigate some major concerns about realizing such a small sized robot. First we introduce the concept of self-assembly and introduce examples both from nature and artificial products. Followed by the main problems in self-assembly which can be seen in various scales, we classify them into four groups - (A) assembly constraint issues, (B) stochastic motion issues, (C) interactions on physical property issues, and (D) engineering issues. Then we show a segregation effect with our developed platform as an example of self-organizing behavior achieved in a distributed manner.

In the area of bipedal locomotion, the spring-loaded inverted pendulum model has been proposed as a unified framework to explain the dynamics of a wide variety of gaits. In this paper, we present an analysis of the mathematical model and its dynamical properties. We use the perspective of hybrid dynamical systems to study the dynamics and define concepts such as partial stability and viability. With this approach, on the one hand, we identify stable and unstable regions of locomotion. On the other hand, we find ways to exploit the unstable regions of locomotion to induce gait transitions at a constant energy regime. Additionally, we show that simple nonconstant angle of attack control policies can render the system almost always stable.

In this paper, we will argue that if we want to understand the function of the brain (or the control in the case of robots), we must understand how thebrain is embedded into the physical system, and how the organism interacts with the real world. While embodiment has often been used in its trivial meaning, i.e. ‘intelligence requires a body’, the concept has deeper and more important implications,concerned with the relation between physical and information (neural, control) processes. A number of case studies are presented to illustrate the concept. These involve animals and robots and are concentrated around locomotion, grasping, and visual perception. A theoretical scheme that can be used to embed the diverse case studies will be presented. Finally, we will establish a link between the low-level sensory-motor processes and cognition. We will present an embodied view on categorization, and propose the concepts of ‘body schema’ and ‘forward models’ as a natural extension of the embodied approach toward first representations.

null

Instability problems regarding biped locomotion increase dramatically, when the points of support are reduced from four to two points. This thesis aims to describe the construction and programming of a biped skating robot, called «Skaty». By trying to implement the principle of cheap design and adapting Skaty to her ecological niche as good as possible, a relatively simple design was developed. In Skaty’s environment, gravity and friction are the dominant forces. Skaty takes more advantage of these forces than fighting against them. Each leg is supported by two passive wheels only. Being equipped with four actors only, Skaty is able to generate slow skating motions in forward direction at an maximum speed of 11 cm ± 10%. No sensors were needed. A natural frequency of 0.9 Hz was predicted, based on a simple physical theory of the skating mechanism. The measured natural frequency at best performance is 1.1 Hz (±10%).

This paper describes the predator and prey robot competition that took place within a robotics class for teachers. The robotics class was part of a degree program that aims at educating upper secondary school teachers of different backgrounds in informatics, a discipline that is not yet a mandatory part of the Swiss school curriculum. The aim of this robot competition was to familiarize the teachers with robotic hardware and software such that they would be able to design their own informatics class syllabus. This paper describes the custom robotic platform used, the competition, its aims and results.

The human hand is one of the most complex structures in the body, being involved in dexterous manipulation and fine sensing. Traditional engineering approaches have mostly attempted to match such complexity in robotics without sufficiently stressing on the underlying mechanisms that its morphology encodes. In this work, we propose an artificial skin able to encode, through its morphology, the tactile sense of a robotic hand, characteristic to slippage events. The underlying layout consists of ridges and allows slippage detection and the quantification of slippage speed. Such encoding of slippage signal becomes suitable for relaying tactile feedback to users in prosthetic applications. This approach emphasizes the importance of exploiting morphology and mechanics in structures for the design of prosthetic interfaces.

Amputees not only lack motor function, but also sensory feedback of the missing limb. It has been shown that lower limb amputees can improve certain gait characteristics when they perceive additional information about the kinematics and kinetics of their prosthetic leg. In this paper, we address the question whether it is feasible to provide centre of pressure location information via electrotactile displays by exploiting the phantom sensation phenomenon, where relative intensity of two electrode pairs is used to encode position between them, creating a single illusory stimulus. Four healthy subjects were asked to identify different locations or movement patterns of the illusory stimulus on a discrete scale under static and dynamic conditions. These stimuli resembled CoP patterns in different locomotor activities. An average recognition accuracy of 73% (std. dev. 17%) was achieved under static conditions, and of 71% (std. dev. 11%) under dynamic conditions. This indicates that the proposed display and mapping can be used to present centre of pressure location, and future work will focus on evaluation with patients.