Monthly Archives: September 2013

Assembly and disassembly of magnetic mobile micro-robots towards deterministic 2-D reconfigurable micro-systems

Authors: Eric Diller, Chytra Pawashe, Steven Floyd, and Metin Sitti.

Source: The International Journal of Robotics Research.

Overall Review:

In this paper, a novel set of permanent magnet modules that are under 1mm in all dimensions is proposed to use in a reconfigurable micro-system. They are capable of locomoting on a 2-D surface. Multiple modules are controlled by using an electrostatic anchoring surface. Both assembling and disassembling are discussed and demonstrated in this paper.

In this paper, two kinds of micro robots are employed: shell-less robots, and shell-based robots.


1. In Section 5.2, “On a polyurethane surface, it is clear that the shell-less Mag-uMods, which contact the surface with MPIP, are exhibiting higher friction coefficients than the shell-based Mag-uMods, which contact the surface with its ALIP shell. This difference would indicate that the polyurethane that binds the metallic powders in ALIP and MPIP is not the exposed material.” What is the meaning of the last sentence? What is the “exposed material”? Why the results imply that the polyurethane is not the “exposed material”?


Agent-based AIS approach for adaptive damage detection in monitoring networks

Author: Bo Chen.

Source: Journal of Network and Computer Applications 33 (2010) 633-645

Overall Review:

This paper presents an agent-based artificial immune system approach for adaptive damage detection in distributed monitoring networks. In this method, a group of autonomous mobile monitoring agents mimic immune cells (such as B-cells), and interact locally with monitoring environment, and respond to emerging problems through simulated immune responses. A scaled bridge model is used to validate the performance of this method. Results are satisfying.

This method can be classified as “Supervised Learning“.


1. In Section 4.2.2, the author discussed two scenarios where the memory cell set will get updated. But what if there is a “bad” memory cell in the set? It has low affinity value with any antigens in his class and thus will never become the “matched memory cell”. Using the proposed method, it will never get replaced.

Answer: This problem will not happen. Because all memory cells are selected from candidate antibodies, which all have high affinity values with one specific kind of antigens. We don’t initialize the memory cell set using randomly generated or other method generated data.

2. Continue with Question 1, what if you have a “bad” antigen at the beginning of your memory cell set generation process? The “bad” antigen will give you a “bad” candidate antibody, and then the “bad” candidate will be add to the memory cell and never get replaced or deleted.

A hybrid immune model for unsupervised structural damage pattern recognition

Authors: Bo Chen and Chuanzhi Zang.

Source: Expert Systems with Applications 38 (2011) 1650-1658

Overall Review:

In this paper, authors present an unsupervised structural damage pattern recognition approach based on the fuzzy clustering and the artificial immune pattern recognition (AIPR). This method has been tested using a benchmark structural proposed by the IASC-ASCE Structural Health Monitoring Task Group. Results show the feasibility of using the hybrid AIPR method for the unsupervised structural damage pattern recognition.


1. For Fig. 3, there are “Meet injection criterion?” and “Meet replacement criterion?” two steps. What is the point of using these two steps?

Answer: The memory cell set needs to have a high diversity. If the candidate memory cell is very different with existing memory cells (measured by rms), then it is “injected” into the memory cell. Otherwise, the candidate cell is compared with existing memory cells. If the candidate cell is “better” (has higher affinity value than the matched memory cell) and it is really close with the matched memory cell, then it replaces the matched memory cell (the total number of memory cells doesn’t change). Or if it is better than the matched memory cell but they are not that close, the candidate memory cell is added to the memory cell set (we have one more memory cell now). To sum up, if the candidate memory cell is quite “different”, or has better “performance”, it will be added into the memory cell set. If there’s an existing memory cell similar with it, this memory cell will get deleted.

2. In Section 2.4.1 “For a training antigen ag, the affinity between the antigen and each antibody ab that is in the same pattern as the antigen is calculated.” How can we know the antigen’s label in advance? Isn’t this an unsupervised learning algorithm?

Answer: This is an unsupervised learning algorithm. A fuzzy clustering algorithm is used to classify all feature vectors into several patterns (using representative point). The antigen’s label is generated by the classification. But we still don’t know the “real label” of this antigen.

3. In Section 3 “In the experimental study, a total of 15 accelerometers were used …, three accelerometers for each level.” This means we have 5 levels for the structural. But the benchmark structural only has 4 levels.

4. In Section 3 “To test the memory cells generated by the immune learning process, the previously created 580 feature vectors were reused with pattern labels. These feature vectors were classified by the memory cells to five clusters.” What is the criterion used for this classification? Is it the nearest neighbor criterion? Is it the Euclidian distance? If yes, is there a threshold? For example, if the unknown antigen’s affinity values with every memory cell are all lower than the threshold, this unknown antigen will not get classified.

5. In Section 4 “The pattern recognition success rate rises rapidly when the MCRT value is greater than 0.85. The reason is that more candidate memory cells are injected into the memory cell set.” A higher MCRT value means less replacement. How can this leads to a higher success rate?

Answer: As explained in Question 1, if we keep other criteria the same, and only change the MCRT to a higher value, less replacement will happen. But the candidate memory cells which do not replace existing memory cells are not abandoned, instead they are injected into the memory cell set. Therefore we actually have more memory cells when the MCRT value is higher.

Neodymium Magnet (NdFeB/NIB/Neo Magnet)

A neodymium magnet (also known as NdFeB, NIB, or Neo magnet) is the most widely used type of rare-earth magnet. It is a permanent magnet made from an alloy of neodymium, iron and boron to form the Nd2Fe14B tetragonal crystalline structure. Neodymium magnets are the strongest type of permanent magnet made. It has been applied to motors in cordless tools, hard disk drives and magnetic fasteners.

In practical, the magnetic properties of neodymium magnets depend on the alloy composition, microstructure, and manufacturing technique employed.

The above information comes from Wikipedia (