Keynote Lectures

Abstract
In this talk I will present some control aspects related to physical Human-Robot Interaction (pHRI). In order to achieve the goal of a safer, dependable, and high-performance collaboration of robots and humans in industrial and professional service tasks, an integrated approach is needed, where mechatronic design, sensory information, on-line task planning, and reactive control issues are combined in a single framework. At lowest level, I will illustrate the control design for lightweight robots with compliant joints, possibly driven by variable stiffness actuation (VSA). In the latter case, simultaneous and decoupled control of both robot motion and compliance is possible. This allows implementing a safe strategy by increasing compliance when the robot is moving fast, so as to reduce the risk of potential injuries in undesired collisions. For dynamically varying and uncertain environments, the typical case in HRI tasks, on-line collision avoidance methods driven by exteroceptive sensors and rapid detection of unavoidable physical contacts using only encoder sensing will be considered. These situations ask for a portfolio of consistent robot reaction strategies in a hierarchy of safety, co-existence, and active collaboration in the human-robot interaction. Some illustrative examples will be given. The presentation will mainly highlight the latest results obtained in the on-going European project SAPHARI, including also recognition of human motion intentions, human-driven kinestethic learning of robot motion, and reactive action generation patterns.

Abstract

In this talk, we present recent results on the stability and robustness properties of transportation networks for various agents' route-choice behavior. We perform the analysis within a dynamical system framework over a directed acyclic graph between a single origin-destination pair. The dynamical system is composed of ordinary differential equations (ODEs), one for every link of the graph. Every ODE is a mass balance equation for the corresponding link, where the inflow term is a function of the agents’ route-choice behavior and the arrival rate at the base node of that link, and the outflow term is a function of the congestion properties of the link.

We propose a novel decision framework, where the drivers combine their historical knowledge about the global congestion levels with real-time local information to make route choice decisions at every node. We show that, if the rate of update of global information is sufficiently slow and if the drivers make route choice decisions cooperatively, then the Wardrop equilibrium is globally asymptotically stable. We then study the resilience of the flow transferring capability of the whole network under disturbances that reduce the flow carrying capacity of the links. In particular, we characterize various margins of resilience of the network with respect to the topology, 'pre-disturbance' equilibrium, and agents' local route-choice behavior. We show that the cooperative route choice behavior is maximally resilient in this setting. We also setup a simple convex optimization problem to find the most resilient 'pre-disturbance' equilibrium for the network and determine link-wise tolls that yield such an equilibrium. Finally, we extend our analysis to link-wise outflow functions that accommodate the possibility of cascaded failures and study the effect of such phenomena on the margins of resilience of the network.

**This work is done in collaboration with Giacomo Como, Ketan Savla, Daron Acemoglu, and Emilio Frazzoli.

Abstract
The aim is to develop advances optimization techniques from the perspective of combinatorial design of automated manufacturing systems. The suggested methodology is based upon advanced line balancing methods and process planning and equipment selection algorithms. The main results are based on exact mathematical programming methods and their intelligent coupling with heuristics and metaheuristics. A decision-aid system based on these methods will be presented. This system is employed for the design of mass production dedicated machining lines as well as for reconfigurable manufacturing systems.

Abstract
The aim is to develop advances optimization techniques from the perspective of combinatorial design of automated manufacturing systems. The suggested methodology is based upon advanced line balancing methods and process planning and equipment selection algorithms. The main results are based on exact mathematical programming methods and their intelligent coupling with heuristics and metaheuristics. A decision-aid system based on these methods will be presented. This system is employed for the design of mass production dedicated machining lines as well as for reconfigurable manufacturing systems.

Abstract
This keynote presents methods and technologies used in guidance, navigation and control of space robotics. It addresses some basic issues related to control of space manipulators and autonomous flying and ground travelling vehicles. Space manipulators such as, Remote Manipulator System (RMS), Space Station Remote manipulator System (SSRMS) and Special Purpose Dextrous Manipulator (SPDMD) were used successfully for variety of space tasks.
Robotic manipulators in Space pose several navigation and control challenges. Robots structures and joints are flexible and highly nonlinear and their designs include several degrees
of freedom (DOF). Each degree of freedom has to be controlled with high accuracy to achieve satisfactory positioning.

Planetary rovers are used in space to explore the planets and moons. There are many types of robotics vehicles. Their main objective is to explore the planets, take samples of ground and rocks at desired location. Often robotics rovers are directed to take images of the surroundings, mapping, as well as, performing other scientific measurements. Accurate localization of autonomous robots in featureless environment is very difficult and often requires advanced methods of guidance and navigation. It is worth noting that existing sensor as gyroscopes, accelerometers and other inertial systems may not work properly when robot is travelling with low speed. Moreover, odometric sensors such as, encoders and tachometers may also not work properly in the hostile environment of outer planets. In such environment, sensor fusion becomes one of the most important technologies used for guidance, localization, and navigation of planetary robotic vehicles. Integrating several measurements and several types of sensors to retrieve accurate data is one of the most successful technologies used in autonomous robotics.

Free – flying and orbiting robots also will play an important role in Space exploration and servicing. De‐orbiting satellites and servicing space floating vehicles on orbit is an important
part of robotics activities in Space. Although, there are not too many successful applications of such types of robots, it is expected that in future they will play much more important role.
General methods of navigation are shown and discussed and their possible application in Space are presented and discussed.