604
be remediated and supported by specialised mining companies. Unawareness of the exact
locations, extent and conditions of such mines can be a great danger to workers.
Scientists of the TU Bergakademie Freiberg investigate technologies in a number of
projects and implement prototypes to support those activities. In the following, the project
ARIDuA, its possible use-cases, requirements and challenges are shown in two examples.
The researchers focus on the practical aspects during the implementation and deployment
phase. In specific one may see that besides the increasing safety underground, robots and
sensors themselves can become a source of danger.
2 PROJECT OVERVIEW
Robotics and IoT technologies are research targets of the ARIDuA-Group (Autonomous
Robots and Internet of Things in underground mining) at TU Bergakademie Freiberg
(Germany). These junior research group, founded in 2017, has an interdisciplinary structure
in order to benefit from synergies through the combination of technology concepts. Six
PhD students from the disciplines of computer science, materials science, mine surveying
and mining, develop interdisciplinary concepts and prototypes with the aim of progressive
digitization and automation in underground mines.
The ESF-funded project (European Social Fund), which runs until mid-2020, aims to
develop an autonomous moving robot for the installation and maintenance of underground
IoT infrastructure (Lösch et al. 2018b). The mining robot “Julius” from the previous project
“Mining-RoX” (Grehl et al. 2017) id used as a platform and will be developed further. Among
other things, a wireless sensor network (WSN) is installed. To further benefit from this structure, mobile sensor box integrate a communication layer and act as nodes within this network, instrumenting the mine itself. These sensor boxes should be able to record relevant
ventilation data using sensors and connect wirelessly both to the Internet and other devices.
In addition, evaluation algorithms for predicting dangerous events using neural networks
based on the acquired sensor data will be developed. In parallel, a classification algorithm for
the detection of ore veins based on optical characteristics will be created (Varga and Grehl
2018), so that all technologies together are a first step to a unmanned mine. The universities
own research and teaching mine “Reiche Zeche” in Freiberg is available as a unique test environment (Mischo 2015). Sections of the mine (depth: 150 m) are used for underground tests
in a real-world environment.
3 POSSIBLE APPLICATION SCENARIOS AND TASKS
Before developing and installing underground wireless networks, neuronal networks, autonomous mining robots or mobile sensor boxes, the question of the purpose for which these
technologies are to be used must be answered. As a boundary condition for the ARIDuA
project, the sites should be limited to underground mines. Within the framework of the
project, the focus will be on mines in Saxony, Germany. Both the research and training mine
“Reiche Zeche” and the abandoned mines in the Ore Mountains (Erzgebirge) are former
ore mines, where mainly silver, non-ferrous metals and uranium were mined. In these mines,
active mining was stopped at the latest in the second half of the 20th century. Today, they are
mostly used for tourism. In total, there are 52 mines open to visitors in Saxony (Sächsisches
Oberbergamt 2016). Other historic mines have been partly abandoned since the Middle Ages
and their existence or condition is poorly known. The only three active underground mines in
Saxony are located in Niederschlag (fluorspar), Seilitz (kaolin) and Hammerunterwiesenthal
(marble) and are managed by small companies (Sächsisches Oberbergamt 2014). While there
is a large number of remediation work underground in old and abondened former mines.
The resulting technologies can be also used in regular operation, for scheduled monitoring or maintenance of the mine. Possible tasks include the automated sensoric measurement
of environmental conditions. In addition to ventilation data such as gas concentrations or
be remediated and supported by specialised mining companies. Unawareness of the exact
locations, extent and conditions of such mines can be a great danger to workers.
Scientists of the TU Bergakademie Freiberg investigate technologies in a number of
projects and implement prototypes to support those activities. In the following, the project
ARIDuA, its possible use-cases, requirements and challenges are shown in two examples.
The researchers focus on the practical aspects during the implementation and deployment
phase. In specific one may see that besides the increasing safety underground, robots and
sensors themselves can become a source of danger.
2 PROJECT OVERVIEW
Robotics and IoT technologies are research targets of the ARIDuA-Group (Autonomous
Robots and Internet of Things in underground mining) at TU Bergakademie Freiberg
(Germany). These junior research group, founded in 2017, has an interdisciplinary structure
in order to benefit from synergies through the combination of technology concepts. Six
PhD students from the disciplines of computer science, materials science, mine surveying
and mining, develop interdisciplinary concepts and prototypes with the aim of progressive
digitization and automation in underground mines.
The ESF-funded project (European Social Fund), which runs until mid-2020, aims to
develop an autonomous moving robot for the installation and maintenance of underground
IoT infrastructure (Lösch et al. 2018b). The mining robot “Julius” from the previous project
“Mining-RoX” (Grehl et al. 2017) id used as a platform and will be developed further. Among
other things, a wireless sensor network (WSN) is installed. To further benefit from this structure, mobile sensor box integrate a communication layer and act as nodes within this network, instrumenting the mine itself. These sensor boxes should be able to record relevant
ventilation data using sensors and connect wirelessly both to the Internet and other devices.
In addition, evaluation algorithms for predicting dangerous events using neural networks
based on the acquired sensor data will be developed. In parallel, a classification algorithm for
the detection of ore veins based on optical characteristics will be created (Varga and Grehl
2018), so that all technologies together are a first step to a unmanned mine. The universities
own research and teaching mine “Reiche Zeche” in Freiberg is available as a unique test environment (Mischo 2015). Sections of the mine (depth: 150 m) are used for underground tests
in a real-world environment.
3 POSSIBLE APPLICATION SCENARIOS AND TASKS
Before developing and installing underground wireless networks, neuronal networks, autonomous mining robots or mobile sensor boxes, the question of the purpose for which these
technologies are to be used must be answered. As a boundary condition for the ARIDuA
project, the sites should be limited to underground mines. Within the framework of the
project, the focus will be on mines in Saxony, Germany. Both the research and training mine
“Reiche Zeche” and the abandoned mines in the Ore Mountains (Erzgebirge) are former
ore mines, where mainly silver, non-ferrous metals and uranium were mined. In these mines,
active mining was stopped at the latest in the second half of the 20th century. Today, they are
mostly used for tourism. In total, there are 52 mines open to visitors in Saxony (Sächsisches
Oberbergamt 2016). Other historic mines have been partly abandoned since the Middle Ages
and their existence or condition is poorly known. The only three active underground mines in
Saxony are located in Niederschlag (fluorspar), Seilitz (kaolin) and Hammerunterwiesenthal
(marble) and are managed by small companies (Sächsisches Oberbergamt 2014). While there
is a large number of remediation work underground in old and abondened former mines.
The resulting technologies can be also used in regular operation, for scheduled monitoring or maintenance of the mine. Possible tasks include the automated sensoric measurement
of environmental conditions. In addition to ventilation data such as gas concentrations or