Cognitive Robotics: Bionic Handling Assistant Inverse Kinematics Study

This research paper delves into the application of inverse kinematics within the context of a Bionic Handling Assistant, a robotic platform inspired by an elephant trunk. It explores the concepts of cognitive robotics, motor learning, and the challenges of controlling flexible robots with infinite degrees of freedom. The paper investigates the use of simulation technologies, particularly C++ and MATLAB, to model and analyze the robot's behavior, addressing issues such as actuation ranges and non-stationary behavior. It also reviews existing literature on continuum robots and exploratory learning of inverse kinematics, aiming to understand the efficiency and limitations of current simulation codes in enhancing the Bionic Handling Assistant's capabilities. The ultimate goal is to contribute to the development of advanced robotic systems that can effectively interact with the environment and humans.

Bionic Handling Assistant
Research Paper
Learning Inverse Kinematics of an Elephant Trunk Simulation
1

Bionic Handling Assistant
Learning Inverse Kinematics of an Elephant Trunk Simulation
Outline
Do Visualization of Bionic Elephant Trunk allows the Efficient Reach ability to
reach the target position?
I Introduction
The concepts of kinematics are widely used in the developing the modern technology for the advanced
robots. . In this series Bionic Handling Assistant is a new robotic platform for developing the modern robots. The
Bionic Handling assistant is developed by the Festo. The Bionic handling assistant is an elephant trunk robot.
Basically the structure of the BHA is manufactured following the concepts of the rapid prototyping methods applied
for 3-D printing. The main task of the BHA is to work as a light weight free moving third hand system. The Bio
tonic handling assistant performs the collaborative activities as well teaching the physical activities of the robot, it
also supports assembly processes and support in handling damageable substances too. This robot is pneumatically
actuated and is manufactured completely with the help of elastic polyamide substance that helps the BHA to become
very flexible and light weight. In robotics the motor learning is very significant in studying the important elements
in robotics as it helps in providing the advance features in the robotics. Currently new generation robots combine the
feature of the mechanical flexibility, elasticity of the material and studying the actuation light weight for the material
used in designing the robotics. According to the study the actuator of the current robots are inspired by the
biological actuator such as Octopus Arm, Elephant Trunk and human biomechanics. That provides the advance
capability for the Robot to have interaction between the robots and the world.
In the artificial systems the analysis is convent ally termed as “motor babbling” where the commands of
the motor are selected in the random manner and results of its behaviour are studied. This theory becomes
irrelevant with respect to the advance robotic machines due to increasing dimension of the sensor motor space. This
required developing more active schemes for supporting the requirement of the modern robotic systems. In this
respect the new technique –“Goal Babbling “approach was introduced. With the help of Goal Babbling approach the
in getting the good results in the world experiments conducted on the non-stationery robot, traditional feedback
control, actuation ranges, drifting sensors and morphological changes in robots were able to study. This also helped
in eliminating the problem of non-stationary and other problems associated with actuation. That provides the
advance capability for the Robot to have interaction between the robots and the world. Most importantly it provides
a great scope of interaction of robot with the human being. But the down side of this feature is that the biologically
inspired design are studied with the help of analytical models for their control and it currently are hardly available,
this fact inspired us to perform the in depth study in this field.
As it is evident that most of the robotic application depends on the task space controller as the main
objective of these controller is to guide the end effectors trajectory of the robotics. As it is evident that these
controllers can act directly on the actuator space a casual mapping between the task space and the actuator space is
2

Bionic Handling Assistant
expected. With the help of the concepts of the Kinematics the derivation of the task space coordinates for the
actuator are derived. Inverse kinematics is extensively used for studying the analytical and machine learning
methods for developing the advance robotics system .It is difficult to develop the model for its model The soft
robots are difficult to develop its model due to its infinite degree of freedom and elastic nature. According to current
study the concepts of the machine learning were found helpful in designing the model for robotics to some extent as
these concepts help to reduce the complexity of designing. In robotics these data are used in studying the kinematics
with developing the kinematics equations. That further help to determine the joint parameters that provide the
specification of the position of the end-effectors for the robot Though study of the kinematics for developing the
advance robotics system provide great support but still there are two major problems associated with it, firstly there
exists an infinite number of possible solutions for the non-inverse homogenous kinematics equation for the
redundant system and secondly the problem of the non –convexity of the solution set.
The software abstraction of the inverse Kinematics has been performed and simulation of the inverse kinematics has
been done with the help of the C++ code. The wonderful results were obtained through this code. This code is
available as open source. It also helps in solving the Bionic handling assistant for solving the control problems in the
robots. This algorithm study also helped in studying the machine learning concepts for the robots. Along with this
the adaption of the MATLAB has also provided a great support in achieving the benefits of kinematics (K. Hosed,
S. Sekimoto, Y. Nishigori, S. Takamuku, and S. Ikemoto, 2012)
II Scope
Cognitive Robotics is an advance field in which advance technologies are incorporated to provide an
important feature in the Robots that they can learn from the environment, from human teachers and from their own
experience. This feature in the in cognitive robots provide the ability of effectively handling the situation in the
environment/The main objective of designing these robots is to provide them capability to react appropriately in the
real life situations. In cognitive technology the main features are expected to be given in the robots are as follows-
 Machine Vision
 Voice Recognition
 Speech Synthesis
 Proximity synthesis
 Pressure sensing
 Texture Sensing
 Programmable motion
 Anticipation and planning
 Imitating particular motion
 Ability to learn mistakes
 Acquisition of knowledge for long term use
 Ability to explore
With the help inverse kinematics concepts in Bionic Handling Assistant the features have been incorporated. Here
in this research the detailed study of the Bionic Handing Assistant will be done. The main objective of this research
Study is to learn about the advance feature of Bionic Handling Assistant and how it can be helpful in designing the
modern robots. Most importantly in this research study the efficiency of the available codes in C++ /MATLAB for
3

Bionic Handling Assistant
simulation in inverse Kinematics for Bionic Handling Assistant will be studied. In c++ these codes are available at
the open source. Along with this the limitation of these codes in providing the efficiency to the Bionic Handling
Assistant` will be explored.
III Research Questions
As discussed in the Scope section there are various elements to perform the study in the cognitive robotics and
application of inverse Kinematics in designing the Bionic Handling Assistant. It is a newly designed robotics
environment and has provided a great support in designing the modern robots too. In this research study following
research questions will be studied-
1. What are cognitive robotics and its impact on modern robotics?
2. What is the application of the inverse kinematics in designing the modern robots?
3. What is Bionic Handling Assistant?
4. What are the technologies that are used in designing the Bionic Handling Assistant?
5. How can simulation technology be used in designing effective Bionic Handling Assistant?
6. What is efficiency of the available codes in C++ /MATLAB for simulation in inverse Kinematics
for Bionic Handling Assistant?
7. What are the limitation of these codes in providing the efficiency to the Bionic Handling
Assistant`?
The above discussed questions will be tried to explore in this research study in the detailed manner.
IV Bionic Handling assistant
Fig1:Bionic Handling assistant
In the first appearance the Bionic Handling Assistant looks like a flexible gripper arm which is designed as an
elephant trunk to work on it. This platform is used as development platform for robotics where wide range of
technologies and components are combined to develop the modern robotic systems. It works in a freely moving,
easy and in flexible manner. It is also safe if there is direct contact between the human being and the machine. It has
safety mechanism in case of accidental collision. It also serves as a multi-technology platform for the simultaneous
development of the electronics, mechanics and software for the advance machines. It has also the feature of image
and speech detection function that makes it more usable in designing the advance robotic systems. It does not
4

Bionic Handling Assistant
require nay programming work or manual operation to detect the voice and image each time. It also provides the
advance gripper fingers that enable the handling of the fragile and differently shaped items without being destroyed
(D. Nguyen-Tuong and J. Peters, 2011)
The most distinguished feature available in the Bionic Handling Assistant is the computing of the end effectors
position with the help of the geometric information about the actuators. This technique is known as Forward
Inverse Kinematics , Along with this Bionic Handling Assistant also holds the different kind of morphology
known as Continuum Kinematics , it provides the capability to Bionic Handling Assistant for having the infinitely
many degrees of freedom . This is obtained by computing the simulation in the Bionic Handling Assistant using the
inverse kinematics concepts.
V Related work-Literature Review
This is one of the most widely conducted research study conducted in the Robotics. In this series the first
research paper that is going to be presented here is the –
1. “Software Abstractions for Simulation and control of a continuum Robot “
Arne Nordmann, Matthias Rolf, Sebastian Wrede
This research paper has wonderfully presented the detail of the application of the bionic in designing of the
continuum Robot. With the help of this paper , I got detailed information about how the Bionic Continuum
Robot is manufactured with the help of elastic polyamide with applying the prototype procedure. It also serves
as a multi-technology platform for the simultaneous development of the electronics, mechanics and software for
the advance machines. It has also the feature of image and speech detection function that makes it more usable
in designing the advance robotic systems. It does not require nay programming work or manual operation to
detect the voice and image each time. It also provides the advance gripper fingers that enable the handling of the
fragile and differently shaped items without being destroyed. This paper has analyzed the abstraction of the
existing Robot control Interface and compliant control architecture which helped to understood application in
more better manner,. This platform provides the facility to support the software modelling and software
architecture mechanism. This paper has further also explored the efficiency of hardware platform used for
manufacturing the Bionic Hand platform.
This paper helped me to understand the cognitive robotics and its impact on modern robotics.
Further this paper also helped me to get details of the technologies that are used in designing the
Bionic Handling Assistant.
2. “Efficient Exploratory Learning of Inverse Kinematics on a Bionic Elephant
Trunk”
5

Bionic Handling Assistant
Matthias Rolf, Jochen J. Steil
This is the wonderful research paper which presents the most important detail about the application of Inverse
Kinematics in the field of Bionic Handling Assistant; this paper has presented the detail of the challenges to be
faced in designing the Bionic Handling Assistant. The biggest challenge according to this paper is handling the
actuation ranges and non-stationary behavior of the object. Here the application of the Goal Babbling approach for
getting the good results in the world experiments conducted on the non-stationery robot, traditional feedback
control, actuation ranges, drifting sensors and morphological changes in robots were shown in the wonderful
manner. Most importantly this paper has presented the re4sults of the simulation conducted with the software
abstraction. With the help of the simulation analysis the impact of non-stationary actuation ranges, drifting sensors
and morphological changes were presented in the detail. Most important feature of this paper is that it has presented
the detailed substantial quantitative evidence of the real-world for obtaining the success of goal-directed
bootstrapping schemes and it also presented the scheme for handling the challenge of non-stationary system
behavior. In robotics the motor learning is very significant in studying the important elements in robotics as it helps
in providing the advance features in the robotics. Currently new generation robots combine the feature of the
mechanical flexibility, elasticity of the material and studying the actuation light weight for the material used in
designing the robotics.
This paper helped me to get details the technologies that are used in designing the Bionic Handling
Assistant.
3. “Qualitative approach for inverse kinematic modeling of a Compact Bionic
Handling Assistant trunk”
M.Achelli , Rochdi Merzouki
This paper presents the detail about the compact Bionic handling assistant (CBHA), it is a continuum
manipulator which works with the help of pneumatic-based actuation and compliant gripper. This
bionic arm is attached to a mobile robot. Further this paper presents the challenges associated with
developing the complex mechanism of the continuum BHA. The key challenge associated with this BHA is
providing the performance control algorithm. This is due to complex mechanical design, hyper redundancy
and presence of uncertainties. There are numerous studies presented to provide the solution but Distal
Supervised Learning (DSL) scheme has shown good results. This paper has presented the qualitative
modelling approach that is based on the neuronal model of the inverse kinematic of CBHA. Along with this
a penalty term constraint is applied to the inverse objective function in the Distal Supervised Learning
(DSL) scheme to select one particular inverse model from the redundancy manifold in it. The explanation
given in the paper has helped me to understand the topic in more analytical way.
6

Bionic Handling Assistant
This paper helped me to understand the application of the inverse kinematics in designing the modern
robots.
4. “Goal babbling for an efficient bootstrapping of inverse models in high dimensions”
Matthias Rolf
This research paper has presented the application of the Goal Babbling scheme in the well structured way.
This is also the main section of study. The paper has highlighted the importance of learning of the high
dimensional motor-system for designing the effective robotic systems. Further the analysis of the
prototypical coordination problem has been investigated. Goal babbling scheme is inspired by the studies
conducted on the infant goal directed attempts. This thesis presents the bootstrapping approach utilized in
the goal babbling scheme. It further highlights the robotics to learn from the inverse model for solving the
coordination problem. Further this paper presents the application of inverse learning in achieving the goal
babbling method. It \further shows the technique to learn the inverse model by fitting the data in it.,
Experiments show that the goal babbling scheme constitutes the positive feedback loop while learning in
the bootstrapping skills.
Table 1: Feed forward control, source: Research gate
This paper is one of most well presented research paper in context of the inverse model application of
the Goal babbling scheme and has helped me to get details about the goal babbling approach in descriptive
manner.
This paper helped to get the details of Goal babbling methods and how can simulation technology be
used in designing effective Bionic Handling Assistant.
VI Methodology applied
7

Bionic Handling Assistant
The purpose of the Research methodology is to understand the scope, rationale and the basic research
approach as well also giving detail of the well researched and practised techniques for conducting this
particular research. Here the study of already published work has been utilized. The intervention approach
of methodology has been applied to study about this subject area. This methodology has been selected due
to its efficiency in the technical, experimental and scientific study. Here as we have to study the
performance of the C++/MATLAB existing codes available for implementation of the Bionic Robots . The
methodology will be as follows-
 The visualization in c++ will be used.
 The existing codes will be run and their efficiency will be evaluated on various parameters.
 The results will be evaluated on the basis of principles of kinematics.
 The results will be documented and will be compared with the results given by the other
observers.
 On the basis of this study the statement of result will be given.
VII Technical Efficacy
Technical Efficacy is a process of performing the experiments in the efficient manner. Here in this study
the existing codes were downloaded from the open source . The software abstraction of the inverse Kinematics has
been performed and simulation of the inverse kinematics has been done with the help of the C++ code. The
wonderful results were obtained through this code. This code is available as open source. It also helps in solving the
Bionic handling assistant for solving the control problems in the robots. These algorithm study also helped in
studying the machine learning concepts for the
Robotics.
Fig 3: sample of C++ code available open-source online for Kinematics: source open-source
The above given snapshot is a part of the code that is needed to compute the BHA’s Kinematics. Here the
efficiency of the c+ /MATLAB code for studying the simulation of this Bionic handling assistant. The inverse
kinematics study of the elephant trunk robot comprises many challenges such as study of the non-stationary
behaviour, high dimensionality, restrictive ranges, actuation ranges and other aspects. Here the main technique used
in exploring the scheme through the –“Online Goal Babbling “technique.
8

Bionic Handling Assistant
In the main section the more detailed analysis of the efficiency of the Online Goal Babbling “technique. For
simulation of the BHA’s Kinematics. The detailed working of the BHA will be presented. Here the main
consideration will be studying the efficiency of the existing MATLAB/C++ code for simulation of Bionic Handling
Assistant and the Goal Babbling algorithm for the detailed study will be presented. The Goal Babbling algorithm
deals with challenge associated with the sensor motor tasks conducted by the robots. The related study in this
subject area will also be presented.
Method and results: I am planning to place a target position in the robots workspace and make the end
effectors of the robot move to that target position. There will be a lot of bending and stretching of the BHA by using
this process.
Experiment 1
Small change in length of actuators can lead to large and direction-wise inhomogeneous changes of the effectors
position. Figure three from the research paper.
Experiment2
Representing three-dimensional workspace description constructed from plain grid by rotating it around five
different angles. The main objective is to learn the inverse model.
VIII Result -Software abstraction of the Elephant Trunk Simulation
The software abstraction of the Goal babbling scheme in providing the support to the BHA is used for
performing the simulation. Here in this approach the codes written in C++ and MATLAB have been
utilized. The codes in C++ are available in open- source. The software abstraction presented the detail of
the existing Robot control interface (RCI) and compliant control architecture (CCA) that is used to
investigate the efficiency of software architecture and software modeling. Here the focus is given on three
challenges to overcome, the first challenge to overcome is enabling reasonable and hierarchical semantic
abstractions of the robot, the second challenge is develop the hardware I/O abstractions for the prototypical
and heterogeneous technical set up and third challenge is to develop the mechanism that realize that it is a
flexible and reusable procedure. In further section the efficiency of c++ and MATLAB codes will be
explored.
9

Bionic Handling Assistant
Fig 6: Simulation of MATLAB/C++
According to the research reports the c++/MATLAB has shown efficient result in presenting the control
with using the goal gabber scheme. The wonderful results were obtained through this code. This code is
available as open source. It also helps in solving the Bionic handling assistant for solving the control
problems in the robots. This algorithm study also helped in studying the machine learning concepts for the
robotics.
Result outcome-
The experiments with the BHA doesn't exists physically. If we consider about the aspects of pressure the codes
does not show result in efficiency.
Non- stationary Actuation Ranges
Centre problem of BHA is that its length changes over time when the pressure is applied.
High pressure bursts and industry has assigned the right pressure for each actuators.
Kinematics learning problems
Reaching for some desired cartesian position x*E R3 with this robot means to find some posture i.e a combination
of lengths q, that returns in an end-effector position x=x*.
I X Conclusion
The main task of the BHA is to serve as a light weight free moving third hand system. The Bio tonic
handling assistant performs the collaborative tasks as well physical teaching of the robot, it also supports assembly
processes and support in handling damageable substances too. This robot is pneumatically actuated and is
manufactured completely with the help of elastic polyamide substance that helps the BHA to become very flexible
and light weight. In robotics the motor learning is very significant in studying the important elements in robotics as
it helps in providing the advance features in the robotics. Currently new generation robots combine the feature of the
mechanical flexibility, elasticity of the material and studying the actuation light weight for the material used in
designing the robotics. According to the study the actuator of the current robots are inspired by the biological
actuator such as Octopus Arm, Elephant Trunk and human biomechanics. That provides the advance capability for
the Robot to have interaction between the robots and the world. The efficiency of C++/MATLAB in simulation is
overwhelming.
10

Bionic Handling Assistant
Referencing
[1]. D. Nguyen-Tuong and J. Peters, “Model learning for robot control: a
survey,” Cognitive Processing, vol. 12, no. 4 ”Special Corner: Cognitive
Robotics”, 2011
[2]. K. Hosoda, S. Sekimoto, Y. Nishigori, S. Takamuku, and S. Ikemoto,
“Anthropomorphic muscular-skeletal robotic upper limb for understand-
ing embodied intelligence,” Advanced Robotics, vol. 26, no. 7, pp. 729–
744, 2012.
[3] R. Martinez-Cantin, M. Lopes, and L. Montesano, “Body schema
acquisition through active learning,” in ICRA, 2010.
[4] M. Rolf, J. J. Steil, and M. Gienger, “Mastering growth while bootstrap-
ping sensorimotor coordination,” in EpiRob, 2010.
[5] K. Neumann, M. Rolf, and J. J. Steil, “Reliable integration of continuous constraints into extreme learning
machines,” Journal of Uncertainty,Fuzziness and Knowledge-Based Systems, no. Special Issue: Int. Symp.Extreme
Learning Machines, 2013,
[6] P. O. Stalph and M. V. Butz, “Learning local linear jacobians for flexible
and adaptive robot arm control,” Genetic Programming and Evolvable
Machines, vol. 13, no. 2, pp. 137–157, 2012.
[7]. U. Sailer, J. R. Flanagan, and R. S. Johansson, “Eye–hand coordination
during learning of a novel visuomotor task,” Journal of Neuroscience,
vol. 25, no. 39, pp. 8833–8842, 2005.
[8] L. Jamone, L. Natale, K. Hashimoto, G. Sandini, and A. Takanishi,
“Learning task space control through goal directed exploration,” in IEEE
Int. Conf. Robotics and Biomimetics (ROBIO), 2011.
[9] P. O. Stalph and M. V. Butz, “Learning local linear jacobians for flexible
and adaptive robot arm control,” Genetic Programming and Evolvable
Machines, vol. 13, no. 2, pp. 137–157, 2012.
[10] C. Hartmann, J. Boedecker, O. Obst, S. Ikemoto, and M. Asada, “Real-
time inverse dynamics learning for musculoskeletal robots based on echo
state gaussian process regression,” in Robotics: Science and Systems
(RSS), 2012.
[11] M. Rolf, J. J. Steil, and M. Gienger, “Mastering growth while bootstrap-
ping sensorimotor coordination,” in EpiRob, 2010.
[12] A. Baranes and P.-Y. Oudeyer, “Active learning of inverse models
with intrinsically motivated goal exploration in robots,” Robotics and
Autonomous Systems, vol. 61, no. 1, pp. 49–73, 2013.
[13] “Deutscher Zukunftspreis (German Future-Award),” 2010. [Online].
Available: http://www.deutscher-zukunftspreis.de/en/content/2010
[14] K. Neumann, M. Rolf, and J. J. Steil, “Reliable integration of continuous
constraints into extreme learning machines,” Journal of Uncertainty,
Fuzziness and Knowledge-Based Systems, no. Special Issue: Int. Symp.
Extreme Learning Machines, 2013, in Press.
[15] T. D. Sanger, “Failure of motor learning for large initial errors,” Neural
Computation, vol. 16, no. 9, 2004.
11

Bionic Handling Assistant
12