The “ ” issue is becoming central in the news mainstream and it will probably re-shape the vehicle market in the next few years. self driving When we think about “ ” and its technical challenges it is always difficult to be up-to-date on the emerging ideas and concepts. self driving It’s easy to accept the wrong inference that, since the web is full of technical info and scientific papers, the knowledge is also readily usable. Unfortunately, are often affected by (and sometimes self-referential) jargon and by the of publications. scientific papers hypertechnical hyperproduction The starting point In order to try to extract some insights regarding this topic, we’ve analyzed several articles published in the database. arXiv ArXiv seems to be a good starting point: it is probably . the most authoritative preprint database and its API is simple and fast Our study is only at the first step but we wanted to share some early findings and we are . glad to receive ideas for the next steps This analysis does not claim to be an exhaustive overview of the topic. It’s only an attempt to of the mass of information. scratch the surface This work is also a chance for us to try some techniques that could be useful to get without the knowledge of an average PHD in each specific topic (anyway having a PHD title isn’t enough to ensure you the success :) ). insights from a corpus of technical documents Technique We’ve tried several approaches and the following is the one that seems to give some promising starting points. - We’ve got a list of papers from the , starting from some generic queries related to the topic “self driving”. arXiv API - We’ve filtered the results topics (i.e. : all the astrophysics papers) removing obvious non-related - We’ve (tf-idf with some tweaks) and created a based on the . transformed the obtained corpus to a matrix graph connections among words - In order to reduce the noise and to extract clearer data, we’ve detected the inside this graph using a . communities hierarchical algorithm The final product of this process is that allows us to explore some of the most promising and their closeness in the papers. a tree n-gram s The last step of this analysis is similar to what the gold digger does with his sieve. We have tried to , through the reading of some related papers and other material that can be found on the web. analyze more deeply some of the n-grams For this small post we’ve selected the , trying to . top n-grams representing each cluster contextualize the underlying topics Topics Topics emerging from the corpus of papers related to self driving - machine learning algorithms, deep convolutional neural, machine learning techniques, deep neural network, computer vision, lane detection This one is the most expected topic because it is related to some of the keywords prevailing in the soft scientific news and in the modern job titles. and are some of the techniques strongly related to the improvement of the effectiveness of the “computer vision_”_. Machine Learning algorithms deep neural nets - iterative consensus clustering, learning method, gaussian process is an interesting Ensemble Method that allows to in order to improve the precision of the classification. Iterative consensus clustering mix the outcome of several algorithms are everywhere in statistics and in physics so it is not a surprise to find them here. Again, in this cluster as well, we find the learning methods that are “monopolizing” the big data analysis. Gaussian Processes - attacker models, malicious users This pair is really interesting because it shows us that the is taken into account inside the scientific community. security issue With the increasing number of interacting systems and of autonomous algorithms and with the constant production of a huge amount of data, the realm of weak points that can be exploited by a malicious user will increase. -autonomous vehicle control, decision control, linear program, game theory, artificial intelligence, autonomous vehicle Interestingly, the (classic) , and the (in its most general sense) are correctly a key topic, separated from machine learning and neural nets. linear programming game theory artificial intelligence - kalman filter, collision avoidance, autonomous navigation is a fast and reliable technique used extensively in robotics in order to . Here it is related to the topic of “ ” that is (obviously) not a secondary issue when it comes to self driving cars! Kalman filter filter out the statistical noise collision avoidance - fuzzy inverse model, learning algorithm, address problem, internet things are another topic strongly related to this kind of problems. The idea of using fuzzy constraint is not new but it is still useful. Fuzzy models More generally, the idea of some kind of or a set of is a classic topic. fuzzy logic fuzzy constraints Here we found also the connection with the (Internet Of Things), another refrain pumped up by several sponsors. IoT - kitti oxford robotcar, camera without relying, complex urban environments, deep semantic segmentation, lane markings, equipped monocular camera This group contains useful information for someone wanting to try to play with algorithms related to the main topic. Kitti oxford robotcar is a reference to the : . huge Oxford RobotCar Dataset http://mrg.robots.ox.ac.uk/the-oxford-robotcar-dataset/ - deep reinforcement learning, markov decision process We expected to find some more often in this analysis. markov technique Markov techniques are broadly used in because they are powerful but, at the same time, are feasible in terms of computational requests. artificial intelligence We find another deep learning technique in this cluster: which is heavily sponsored nowadays. deep learning - kinodynamic motion planning, finite elements, constrained optimization problem This group of words in really interesting to us because we’ve never heard anything about (our ignorance). It’s nice because it sounds like technobubble from Star Trek and so it reminds us of the 90s. kinodynamic motion planning After a brief study now we know that it is a class of problems for which , together with such as avoiding obstacles. velocity, acceleration, and force/torque bounds must be satisfied kinematic constraints Conclusions Like any well financed and also much exposed target, the “ ” issue is a collector of . self driving interesting new ideas (and also of boring and recycled ones) This topic is particularly and we are sure that in the near future some interesting studies will flood in other sectors. multifaceted It will also be a huge task to s in order to embrace this change of paradigm. In this world where the market of science and tech seems to be self sustained, the target of (not pure science) seems to be the real difficult task. align laws and regulation keeping tech bound to the needs of humanity Who We are , a data company that provides analyses on politics, society and institutions. Elif Lab We are also working on a project called that analyzes various data coming from the European institutions. ThinkingAbout.EU Contact us on our website or send a mail to www.eliflab.com info@eliflab.com
