Autonomous miniature vehicles are widely used to test many types of algorithms and simulate the driving behavior as in the real world. In the last decade, several approaches have been proposed for obstacle avoidance in self-driving vehicles. In this paper, we present our autonomous vehicle implementation which gives a solid solution to this problem in an efficient manner, using a Raspberry Pi and a LIDAR module for indoor navigation. Our vehicle is capable of navigating in an unknown environment while avoiding obstacles. Unlike other implementations, our vehicle doesn't use Computer Vision (CV) techniques for obstacle detection but only a single LIDAR sensor and thus can safely navigate in low luminosity enviroments. We perfomed various experiments and verified the success of our implementation.

Recycling is vital for a sustainable and clean environment. Developed and developing countries are both facing the problem of solid management waste and recycling issues. Waste classification is a good solution to separate the waste from the recycle materials. In this work, we propose a cloud based classification algorithm for automated machines in recycling factories using machine learning. We trained an efficient MobileNet model, able to classify five different types of waste. The inference can be performed in real-time on a cloud server. Various techniques are described and used in order to improve the classification accuracy, such as data augmentation and hyper-parameter tuning. Multiple industrial stations are supported and interconnected via custom data transmission protocols, along with security features. Experimental results indicated that our solution can achieve excellent performance with 96.57% accuracy utilizing a cloud server.

In recent years, the production of municipal solid waste has constantly been increasing. Recycling is becoming more and more important, as it is the only way that we can have a clean and sustainable environment. Recycling, however, is a process that is not fully automated; large volumes of waste materials need to be processed manually. New and novel techniques have to be implemented in order to manage the increased volume of waste materials at recycling factories. In this paper, we propose a novel methodology that can identify common waste materials as they are being processed on a moving belt in waste collection facilities. An efficient waste material detection and classification system is proposed, which can be used in real integrated solid waste management systems. This system is based on a convolutional neural network and is trained using a custom dataset of images, taken on site from actual moving belts in waste collection facilities. The experimental results indicate that the proposed system can outperform existing algorithms found in the literature in real-world conditions, with 92.43% accuracy.

With the advent and rapid growth of automation, unmanned ground vehicles (UGVs) have emerged as a crucial technology, with applications spanning various domains, from agriculture to surveillance, logistics, and military operations. Alongside this surge in the utilization of robotics, novel complications inevitably emerge, posing intriguing questions and challenges to the academic and technological sectors. One such pressing challenge is the coverage path planning (CPP) problem, particularly the notion of optimizing UGV energy utilization during path planning, a significant yet relatively unexplored aspect within the research landscape. While numerous studies have proposed solutions to CPP with a single UGV, the introduction of multiple UGVs within a single environment reveals a unique set of challenges. A paramount concern in multi-UGV CPP is the effective allocation and division of the area among the UGVs. To address this issue, we propose an innovative approach that first segments the area into multiple subareas, which are then allocated to individual UGVs. Our methodology employs fine-tuned spanning trees to minimize the number of turns during navigation, resulting in more efficient and energy-aware coverage paths. As opposed to existing research focusing on models that allocate without optimization, our model utilizes a terrain-aware cost function, and an adaptive path replanning module, leading to a more flexible, effective, and energy-efficient path-planning solution. A series of simulations demonstrated the robustness and efficacy of our approach, highlighting its potential to significantly improve UGV endurance and mission effectiveness, even in challenging terrain conditions. The proposed solution provides a substantial contribution to the field of UGV path planning, addressing a crucial gap and enhancing the body of knowledge surrounding energy-efficient CPP for multi-UGV scenarios.

It is evident that over the last years, the usage of robotics in warehouses has been rapidly increasing. The usage of robot vehicles in storage facilities has resulted in increased efficiency and improved productivity levels. The robots, however, are only as efficient as the algorithms that govern them. Many researchers have attempted to improve the efficiency of industrial robots by improving on the internal routing of a warehouse, or by finding the best locations for charging power stations. Because of the popularity of the problem, many research works can be found in the literature regarding warehouse routing. The majority of these algorithms found in the literature, however, are statically designed and cannot handle multi-robot situations, especially when robots have different characteristics. The proposed algorithm of this paper attempts to give the following solution to this issue: utilizing more than one robot simultaneously to allocate tasks and tailor the navigation path of each robot based on its characteristics, such as its speed, type and current location within the warehouse so as to minimize the task delivery timing. Moreover, the algorithm finds the optimal location for the placement of power stations. We evaluated the proposed methodology in a synthetic realistic environment and demonstrated that the algorithm is capable of finding an improved solution within a realistic time frame.

Due to the Earth’s population rapid growth along with the modern lifestyle the urban waste constantly increases. People consume more and the products are designed to have shorter lifespans. Recycling is the only way to make a sustainable environment. The process of recycling requires the separation of waste materials, which is a time consuming procedure. However, most of the proposed research works found in literature are neither budget-friendly nor effective to be practical in real world applications. In this paper, we propose a solution: a low-cost and effective Smart Recycling Bin that utilizes the power of cloud to assist with waste classification. A centralized Information System (IS) collects measurements from smart bins that are deployed all around the city and classifies the waste of each bin using Artificial Intelligence and neural networks. Our implementation is capable of classifying different types of waste with an accuracy of 93.4% while keeping deployment cost and power consumption very low.