Future Intelligent Machines

Transcription

1 Future Intelligent Machines TKK GIM research institute

2 Content of the talk Introductory remarks Intelligent machines Subsystems technology and modularity Robots and biology Robots in homes

3 Introductory remarks Mechanical engineering led industrial revolution about 250 years ago. Mechanization of work started then and led to amazing increase in the effectiveness of work until automation started about 60 years ago. Automation has increased the effectiveness, but particularly freed people from slavery of machines by increasing machines own intelligence. There are still much room for automation in the society, especially in non-factory applications, like making the machines use by people or the environment where we live, more smart.

4 Intelligent machines what they are? Robotic machines or robots that replace people in tasks, which are monotonous or which do not any more fulfill the minimum requirements for human work or operate in an environment unfriendly for humans or offer help or entertainment services to people in their homes or public places

5 Challenges for technology - Task scenarios often include an environment, which is not well structured, not a priori known in details and may be changing with time. - In most cases robots must be mobile, because tasks cannot be moved to the robots, but robots must move to the tasks. - Often high autonomy is needed meaning that everything, including energy, must be carried on-board. Requirements are similar to biological species.

6 Terminology Field robotics refers to outdoor applications. Many (but not all) applications are built using existing working machines Service robotics refers to mainly indoor applications, but sometime it is also used for light outdoor applications. In many cases service robots are intended to consumer markets for use of common people.

7 Situation in commercial markets Unlike industrial robots service robots as new products has not yet notable markets However, there is a big market potential. Most of the R&D investment world wide in robotics is today in this new area. On the other hand, work machine industry is evolving towards intelligent machines i.e. field robots. These markets are existing and huge.

8

9 Main application fields Working machine technology for farming, construction and utization of natural resources. Exploration of planetary system ( moon, planets, asteroids). Undersea operations Services: home services, entertainment, disabled support, replacing manpower in human-intensive nonindustrial work in offices, hospitals, public places, hostile urban environment (cleaning streets, sewers etc). Military technology Rescue operations and katastrophe aftercare Dismanteling nuclear plants

10 Today s business: Mobile working machines Significant role in Finnish economy 27 biggest exporting companies: turnover 4.4 Mrd eur (2006) export 70% employ more than people Today s newest product development transformes classical manually controlled machines gradually intelligent, autonomous or remote tele-controllable machines.

11 Tomorrow s business: Service robotics Robotics research priority in this area already the last 15 years. Market weight point is gradually moving from industrial robotics to service robotics. On the other hand, only few succeeders this far Roomba from irobot has been sold globally already 2 million pieces. Space and military applications speed up advancement at the moment. Many applications require high machine intelligence, because the robots have to operate in disorganize and badly known environments.

12 Examples Following slides and videos show examples of different field and service robots. There exist much more of them today in web, so you are encouraged to explore web to find more and different type of robots. F&S robots have a great variety of appearance depending their utility area. One of the best site to look different research projects is on-lineexpo 22xx site Hall 5: rsity.htm

13 Today s forest machines are computerized and quite ready to be roboticed Harvester machine - cuts trees and debrach them Forwarder machine - transports trees to roadside

14 Mechant and Plustech walking harvester machine Development done in TKK and Plustech company : Six leg heavy walking machine technology Practical target: harvester machine having mild impact on forest terrain Video

15 Semiautonomous and teleoperated robotic vehicles Mobile robot test-bead constructed from a standard ATV Full scale loader machine instrumented for teleoperation tests Practical target: Remove driver from machine to remote location AutoMine Arska

16 Future multi-purpose service robots P3 (today Asimo) WorkPartner Weight 210 kg Max vel. < 1m/s Work weight 5 kgf/hand Payload? Opration time 15 min kg 2 3m/s 5 kgf/hand 60 kg 4-5 hour (2 liter petrol)

17 Subsystems of F&S robots All F&S robots are systems, which consist of subsystems There is no standardized way to define the subsystems, but the following is de facto standard. The subsystems may consist of HW or SW or both. The cropping attribute is the functionality.

18 Subsystems and their tasks: Power and energy system Motion system Work tool system (includes possible manipulator, other tools and their lower level controllers) Motion control system Perception and sensing system Navigation system Motion and action planning system (mission control) Human-robot interface (intercommunication)

19 Academy of Finland Centre of Excellence in Generic Intelligent Machines Reseach

20 GIM is joint venture of two host organizations Automation Technology Laboratory ATL Founded in 1985; director prof. Halme Belongs to the Department of Automation and Systems Engineering at TKK Institute of Hydraulics and Automation IHA Founded in 1988; director prof. Vilenius Belongs to the Department of Intelligent Hydraulics and Automation at TUT GIM was founded by a common declaration and officially by Finnish Academy

21 General research goal The focus of GIM covers mobile machines, intended to some useful work, from ordinary working machines to futuristic service robots experienced this far only in Star War movies. Vision: The future development of those machines will take place under a generic framework, which opens up novel possibilities to create new mechatronic machines through a modular design and manufacturing. The merging properties are the generic subsystem architecture and physical modular structure that such future machines have. Mission: To bring about unique new research results in the methodology and technology of the subsystems infrastructure that support practical long term development of the generic intelligent machines.

22 Research agenda RP 1 Internal goal assesment and training RP 2 Modularity of generic machines RP 3-4 Power and energy systems RP 5 Control Architecture RP 6 Perception and navigation system RP 7 Human-Robotic Machine Interaction RP 8 Motion Systems RP 9 INTEGRATOR infrastructure for experimentation and validation of reseach results in multi-machine environment

23 Integrator project Integrate machines and people trough Internet Goal is to demostrate Future worksite concept First version of GIMnet already exists and enables to use machines located in Tampere from Espoo Local controls maintenance exceptions Worksite Machines Human workers/ local operators i n t e r n e t GIM- Architecture communication common presence databases Remote control room(s) Model Operator Production management Virtual environment Task design tools

24 Research platform, where all subsystems have been implemented in a non-trivial way WorkPartner developed by TKK Automation Technology Laboratory ( ) Humanoid (centaur-like) robot with two hand manipulator Hybrid locomotion system - wheel driving mode - walking mode - rolking mode (walking on the wheels) Hybrid power system Multimodal HRI based on cognition and interaction Designed for outdoor use

25 GPS-antenna Leg-control unit CCD-camera Navigation unit Main computer Laser pointer Manipulator Engine Body joint Elbow-controller Laser scanner

26 General progress in subsystems technology There has been tremendous progress within last years in several subsystem technology, specifically in - motion control - navigation Still much progress is needed both in methods and technology, specifically in - perception - motion and action planning (AI) - HRI

27 Robots and biology Many problems in robotics can be solved keeping in mind biology and neuro science (bionics, biomimetics): - Improving mechanical design and energy efficiency. Utilization of the energy stored in the environment <-> animal physiology and anatomy - Improving mobility, motion control and coordination of task execution <-> neuro-motoric systems - Improving sensing and environment perception <-> biological sensing and perception system and related neural structures - Designing cognitive interfaces between robots and human users <-> animal and human cognition mechanisms

28 Robotics in homes Robotics is coming as a part of home automation, which in turn is a part of development of building automation When and in what form this will happen is still shrounded in mystery It is a big potential market, however. Challenges in R&D are related in one hand price sensitivity and on the other hand very to the demanding environment that home is. TKK/GIM Ceilbot student innovation project tries a very different solution making the robot use ceiling instead of floor.

29 TKK Ceilbot a big challenge for students to innovate new solutions for home robot using ceiling instead of floor.