Search results

The purpose of this paper is to present the methodology and the results on the design and development of an autonomous, golf ball picking robot, for driving ranges.

The strategy followed to develop a commercial product is presented, based on prior identification requirements, which consist of picking up golf balls on a driving range in a safe and efficient way.

A fully working prototype robot has been developed. It uses two driving wheels and a third cast wheel, and pushes a standard gang which collects the balls from the ground. A hybrid information system was implemented in order to provide a statistically relevant prediction of golf balls location, to optimize the path the robot has to follow in order to reduce time and cost. Autonomous navigation was developed and tested on a simulation environment.

Preliminary results showed that the new path planning algorithm Twin‐RRT* is able to form closed loop trajectories and improve the result over time. Kinematic constraints were already taken into account on the algorithm. This sampling based algorithm has potential usage in solving other TPP (Travelling Purchaser Problem) related problems.

The prototype feasibility is being tested in real driving ranges. It has autonomy of up to 8 h per day. It is capable of collecting up to 1,200 balls in one single journey. It weighs 130 kg and is capable of climbing slopes of up to 22°. The maximum speed is 8 km/h and the robot takes 140 min to completely sweep a 25,000 m² field at 7.2 km/h (2 m/s) average speed.

There are about 30,000 golf practice fields, of which 18,000 are located in the USA and Canada. In some countries the golf industry represents more than 15 per cent of tourism GNP. In a typical practice field, about 10,000 balls have to be picked up every day.

An important contribution of this paper is the algorithm for path planning in order to optimize the ball pick up task, reducing time and cost. There are two patents are pending concerning the technological novelties of this work.

The purpose of this paper is to characterise the TWIN-RRT* algorithm which solves a motion planning problem in which an agent has multiple possible targets where none of them is compulsory and retrieves feasible, “low cost”, asymptotically optimal and probabilistically complete paths. The TWIN-RRT* algorithm solves path planning problems for both holonomic and non-holonomic robots with or without kinematic constraints in a 2D environment.

It was designed to work equally well with higher degree of freedom agents in different applications. It provides a practical implementation of feasible and fast planning, namely where a closed loop is required. Initial and final configurations are allowed to be exactly the same.

The TWIN-RRT* algorithm computes an efficient path for a single agent towards multiple targets where none of them is mandatory. It inherits the low computational cost, probabilistic completeness and asymptotical optimality from RRT*.

It uses efficiency as cost function, which can be adjusted to the requirements of any given application. TWIN-RRT also shows compliance with kinematic constraints.

The practical application where this work has been used consists of an autonomous mobile robot that picks up golf balls in a driving range. The multiple targets are the golf balls and the optimum path is a requirement to reduce the time and energy to refill as quickly as possible the balls dispensing machine.

The new random sampling algorithm – TWIN-RRT* – is able to generate feasible efficient paths towards multiple targets retrieving closed-loop paths starting and finishing at the same configuration.

The author argues that we must stop and take a look at what our insistence on human labour as the basis of our society is doing to us, and begin to search for possible alternatives. We need the vision and the courage to aim for the highest level of technology attainable for the widest possible use in both industry and services. We need financial arrangements that will encourage people to invent themselves out of work. Our goal, the article argues, must be the reduction of human labour to the greatest extent possible, to free people for more enjoyable, creative, human activities.

For intelligent robots in a multi‐agent system communication is essential for cooperative behavior. Here we describe the explicit communication between individual robots acting as team members of a RoboCup team playing soccer. The robots are based on the EyeBot platform. An overview of communication systems being published and a discussion of their advantages and drawbacks is followed by an introduction into multi‐agent systems and the problems we faced applying them to the task of playing soccer. Then we describe the wireless communication network in detail including the EyeBot platform, message structures, self‐configuration and error recovery. The communication allows transmission of messages between individuals, broadcasts and communication with a remote computer workstation. The communication system is a layer beneath the multi‐robot console, which is the user interface, and above the EyeBot hardware.

IN RECENT issues (April, Stopwatch and Book Reviews, June) we referred to the possibility of a factory where the sole human was a person pressing buttons at a control centre. All else was done by automation.

Nissan starts to use robots in car assembly, while Toyota keeps automation to the minimum in latest engine factory

The purpose of this paper is to review the International Manufacturing Technology Show in Chicago with emphasis on innovations in applying automation to manufacturing and assembly.

In‐depth interviews with exhibitors of automated assembly and manufacturing technology.

Building of production machinery continues to move toward offering automated assembly or production cells and away from building single purpose metal cutting machinery. Robots are married to machine tools for much more than just tending.

The paper shows that users in almost any manufacturing realm will find that automated manufacturing technologies are now addressing all types of production requirements. No longer it is necessary to think that only million off part runs can be produced in an automated manner. Machine tool builders are including testing, finishing, marking, assembly and other secondary operations within the basic machining unit for a wide range of production volumes. Automation can even be applied to “made to order” type production.

IN a recent Money programme on BBC TV Sir Monty Finniston was seen advocating the need for greater investment in industry as opposed to money being sent abroad to be invested either in Japanese industry or American real estate.