Sandor Veres

Sandor Veres

Sandor M Veres is professor of Autonomous Control Systems at the Department of Automatic Control and Systems Engineering at the University of Sheffield.

Location Amy Johnson Building, Portobello Street, Sheffield

Activity

  • Tom. You are right. Please note that that I was esrlier saying spatial temporal modelling. Not only coordinates...

  • Interesting comments. There is already discussion emerging in Britain (and in other countries) on societal and wealth-distribution change needed when we use robots in large numbers. The concept of "universal basic income" has been raised, which points towards innovative thinking. In my view we always will have creative areas where robots will not and cannot...

  • I believe we still underestimate the capabilities of the human brain, plenty of scope for better work than "driving a lorry".

  • Great comments. Research is ongoing and we will see how far we will be able to get with learning to drive by machines, watch out for the news.

  • We cannot overdo safety, this is a good suggestion but maybe impractical to ware special west in places where visitors arrive. Public should however be made ware of such zones and sign posts should indicate them. This is in addition to making autonomous cars safe without these additional warnings. We cannot be careful enough.

  • Robotics industry is held back by insurance challenges of autonomous robots, it holds back development and sales. They will however always go as far as they can can with the advanced autonomous features, as long as those do not endanger their profits.

  • Great comments. Robotic guides are difficult to make even nearly as good as human ones. Slowness and inappropriate reactions can put off visitors. It is a challenge to make these robots fast, less prone to mistakes, and entertaining enough for humans to like them.

  • One of the greatest benefits of autonomous cars and monitoring of traffic, which the public will appreciate more than anything else, will be the significant reduction in the number of accidents on our roads in the near future.

  • Great comments. Would add that new laws and legislation will be needed for sensing and decision making processes of autonomous robots in the near future. Itself the procedure of technical certification needs to legislated, which is either too slow or unavailable today. Transparency is needed on this at international level.

  • Apart from detecting other "agents" and objects (pedestrians, cars, balls rolling across the road, roadwork cones, etc. ), the autonomous car also needs a reliable guess of the intentions of other agents and the physics of objects in terms of their behaviour and dangers. In fact it almost needs the judgement of an educated and experienced human driver. This...

  • Great application opportunity, a combination of fixed cameras and autonomous observations and herding by self-charging drones from time to time, could perhaps
    do the job autonomously...

  • I find the comments below on future drone use fascinating and very imaginative, pleasure reading them...

  • There have been EU and other research projects following this approach, their success is yet to be seen in applications...

  • Tom, ... Interesting comment. But don't you think that part of self-awarness is spatial-temporal modelling of the world and where the robot is in it in space and time? I do not see it a separate issue, though I would not call any of today's robots self-aware.

  • Indeed, decision making of robots is most reliable when the robot puts the current situation in the context of the past...

  • Thanks, we will look into this.

  • Robotic research constantly borrows ideas from animals and humans, we try to copy the fascinating capabilities, we are very far from success but initial work started, please see later parts of thiss course.

  • Current revolution of robot development aims at autonomous decision making by robots. When in use in the futute, these may require reallocating staff to higher levels of creativity, where robots will not work yet. No one needs to loose their jobs due to robots with autonomous decisions.

  • Technology development often takes routes incrementally built by market demand, this supports that mobile phones and exoskeletons can be important part of future robotics. Another incremental and market led innovation process is the appearance of ever better autopilots in UAVs , cars and boats. These will become gradually autonomous for increasingly longer...

  • Good comment, future certification of autonomous robots will need to measure sensor reliability and verify that robot decisions take into account uncertainty in sensing... This is short answer only, this topic is huge.