Opendata, web and dolomites

Report

Teaser, summary, work performed and final results

Periodic Reporting for period 1 - ROPOD (Ultra-flat, ultra-flexible, cost-effective robotic pods for handling legacy in logistics)

Teaser

The market for automatically guided vehicles (AGVs) in logistic applications is growing rather slowly in spite of the market potential that has been forecasted. There are markets and applications, which literally cry for automation by AGVs not only for economic, but also for...

Summary

The market for automatically guided vehicles (AGVs) in logistic applications is growing rather slowly in
spite of the market potential that has been forecasted. There are markets and applications, which literally
cry for automation by AGVs not only for economic, but also for socials reasons: hospitals and care facilities.
Irrespective of the burning need there a major barriers, however, which prevent such automation. Two of
them are cost and legacy in existing logistic solutions. Today’s hospital logistic robots are bulky, heavy,
and with a price of 50+ KEUR for a single vehicle they are very expensive. At the same time these AGVs
are often highly specialised and can only deal with few containers or supply carts. In a typical logistic
environment there are however, dozens of different containers and such carts. In a nutshell there is a
bad need for AGVs, which are low-cost and can deal with a great variety of legacy.

The general objective of this proposal is to develop and implement a disruptive concept for AGVs that
lowers the still existing barrier in logistics by offering
• cost-effective, automated or semi-automated indoor transportation of goods,
• while coping with existing legacy in terms of size, shape, and weight of goods and containers,
• without imposing disruptive changes in existing logistic solutions, such as rebuilding entire warehouses
or switching to new containers or storage technology.

We will put an equal emphasis on cost-effective as well as on human-friendly automation of logistic tasks.
While cost-effectiveness shall be achieved by preferably using and adapting technology designed for
mass-markets, human-friendliness shall be achieved by equipping the AGVs with a (semi-) autonomous
shared control mode, in which the robot serves as a force amplifier for the human user and thereby
reduces the physical strain on the user.

Technology development takes place in the context of two use-cases: a) the automatic transportation of
supply carts and b) collaborative transportation of sick-beds. In the first use-case ropods will drive fully
autonomously to the parking position of supply carts, pick them up and move them safely to their target
positions. In the second use case the ropods act as force amplifier, where the human operator is still in
charge for manoeuvring the ropod together with the payload, but can do this with a minimal force applied
to a haptic interface. By exerting force and momentum to the haptic interface the human operator can
then navigate the robots with the sickbed with a single hand.

Work performed

Work in the first reporting period was mainly devoted to development of the various hardware and
software components towards a Technology Readiness Level 4 (Small scale “ugly” prototype built in a
laboratory environment). The main focus of the hardware design was on the development of what we
called the Smart Wheel. This is a highly compact power unit, which includes high torque, high velocity,
small direct drive hub wheels adapted from the electro scooter from the consumer market. It further
comprises a dual-channel motor controller circuit and three magnetic encoders with a resolution
of 19 bits, one for each motor and one for the yaw axis. The direct drive gear together with the high-
resolution encoders and the pivot-mounted suspension are the key ingredient for the compliance of the
power unit which is in turn key enabler for the design of safe robots. Each Smart Wheel has a height of
15 cm at a wheel diameter of 10 cm, a payload of 200 kg at a maximum velocity of 20 km/h.
The development of the Smart Wheel has reached TRL 5. Based on the first prototypes of the Smart
Wheels we further designed a first version of the Smart Vehicle, a quasi-omnidirectional vehicle which
can carry a payload of up to 800 kg. It will be equipped with a docking mechanism that can handle a
variety of the supply carts. We further designed first prototypes of so-called the Sensor Cubes.
Sensor cubes are building blocks of small sensor networks distributed over a single vehicle or a
whole flock of vehicles. They comprise a variety of sensing modalities such a range sensors, IMUs, vision,
3D sensors, or microwave radar.

The main focus of the software design was the layout of the overall software system architecture and the
development of algorithmic functionalities on TRL4. Work involved the selection of an overall conceptual
framework – the Industrial Internet Reference Architecture (IIRA) was adopted and implemented – the setup
of the communication infrastructure within the overall ROPOD system and the integration of the hospital
infrastructure.

The design of algorithmic functionalities such as the development of a fleet management system, a task planner
and scheduler has started with the beginning of Year 2 and is work in progress. TRL4 prototypes that are able
to handle the first use case with a limited number of ropod have been implemented. The same holds true
for the functionalities related to the sensor cubes and the adhoc sensor network. Work included algorithms
for the initialization and self-calibration of the sensor network, sensor data processing and world modelling.
We have decided to use the Open Street Map (OSM) data format and expand and adapt it so that it can be
used for path planning and navigation in large indoor facilities.

Work on basic motion planning functionality such as low-level motion control and motion planning for a
single ropod in a time-varying environment is in progress. First prototypes are available and are currently
integrated in a first prototype of the first use-case. First concepts for a black box were developed in order
to support remote monitoring and diagnosing.

The implementation of the first use-case, automatic transportation of supply carts, has started in the
second quarter of Year 2, involving two weeks of intensive testing of single functionalities and integration
into a first overall prototype of the ROPOD system. These first reality checks have revealed a number of
shortcomings of technology which is state of the art in academia but reaches its limits very quickly in
reality. The consortium is currently addressing those real world issues. This first use case will be subjected
to a more or less regular field tests – several hours on several days per week – before the end of Year 2.
This field test shall demonstrate the socio-economic use of the technology developed in ROPOD so far.

Final results

A main result, which goes clearly beyond the state of the art in mechatronics and has a significant economic
potential, is the development of the Smart Wheel. It enables a technical as well as an economic break-through.
In technical terms we have developed a powerful drive-unit based on components, which originate in the
consumer market. With a payload of 200 kg, a maximum velocity of 20 km/h, and its inherent back-drivability it
outperforms existing drive-units on the market. The twin-wheel based differential kinematics allows an arbitrary
deployment of a set of Smart Wheels under every potential vehicle frame. With this we can shake off the “one
size fits it all” design philosophy that has clearly limited the spread of AGVs, and offer flexible, modular,
customisable solutions for the automatic transportation of large payloads, at a fraction of the cost of todays
AGVs. Automation of hospital logistics becomes affordable and human-friendly.

Website & more info

More info: http://www.ropod.org.