Agile Soft Systems has been providing software solutions in the logistics domain for over a decade. With expertise in Modern Application Development, Agile Soft Systems has been renowned in improving efficiency in fleet management by leveraging cutting edge technologies. Our client, a Fleet Management solutions company, has been a reputed name in the sector, providing end to end solutions to their clients all over the country. They've had a successful track record of partnering and serving companies with small and mid size fleets, especially in the cold storage transportation domain, mainly dairy.

Post pandemic, the market has witnessed a major shift in the technology with the rise of AI, ML and Edge computing, and has had early adopters of such technologies gain a significant share of the market. Our client hadn't revamped their system yet, and was facing fierce competition from companies, old and new. They recently went through a grueling experience trying to retain an old partnership who had been with them since their inception. They had to take immediate action or they would have to suffer dire business consequences.

The client reached out to us to reinvent their existing system with modern technologies to keep them in the forefront of developments in the field. Moreover, revolutionizing their tech was initiated also because the company cared for the problems of their end clients: the companies who had fleets to manage, and their corresponding customers. By partnering with Agile Soft Systems, they were determined to provide solutions to improve safety, efficiency and business advantage for their existing and future partners.

The Business Problem

The system our client was using, relied on legacy architecture and technologies. The data captured from the sensors on the vehicles, was sent to the central server for processing. The software solution wasn't designed to provide real-time assistance to drivers, and the fleet managers were unable to take measures to curb problems, who could only act after the occurrence of an undesired event.

Cold chain monitoring:

The legacy monitoring system was not capable of providing accurate and continuous temperature data to the drivers, risking the quality and safety of perishable goods during transportation. Revolutionizing this aspect was crucial, as a fair portion of the client's customers were into milk transportation. The lack of real-time route optimization too, led to incurred cost and losses to the fleet companies.

Predictive maintenance:

Lack of predictive maintenance led to a higher risk of unexpected breakdowns and increased maintenance costs over time. As the data from the sensors which was sent to a centralized server was accessed by the fleet managers alone, and when left unreported to the drivers, resulted in unplanned maintenance, downtime and poor customer service.

Driver Safety:

The earlier system had a dashboard camera with a view of the road. The recordings from this camera was used as evidence in case of an unfortunate event. The equipment on the vehicle lacked sophisticated driver monitoring features, making it challenging to identify and address unsafe driving practices, potentially leading to increased accidents and liabilities.

Workforce related problems:

The milk transportation industry, which requires seven day week and holiday workdays, is already struggling to maintain a diligent workforce. And the attrition problem is even more severe with fleet management companies who haven't upgraded their transportation system with the latest technologies — the drivers prefer companies with advanced technologies ensuring less responsibilities and additional safety.

Without technological assistance, the drivers faced cognitive load in their day to day activities. In that, they had to be mindful of tasks apart from their core activity, which is driving. If the drivers were found to be at fault, they were either replaced, or taken off the fleet and sent into retraining. This led to bad employee experience and low morale.

Technological Problem

Besides the business problems faced by the client, there were some technological challenges that were identified by Agile Soft Systems. You'll see, how technological limitations were in fact restricting the improvement of the fleet management system.

UX

The legacy system had limited features and the user experience was poor. The fleet managers accessed the system via an interface which had few essential features — GPS monitoring system showed the current location of each vehicle in transit, and a dashboard displayed vehicles which were due for maintenance. There was also a lack of custom communication system between the fleet managers and the drivers. They had to rely on traditional and ineffective modes of communication such as text messages and calls. This process led to discrepancies.

Software and Architectural limitations

The system was designed to send the sensor data to a centralized server for further computation. The drivers were left out of the application's communication loop. They had to rely on common phone apps, for route optimization and for setting up reminders for truck maintenance. Also, the drivers used traditional modes of communication such as text messages and phone calls to communicate with fleet managers and end customers.

Moreover, the software did not use AI and ML algorithms, hence there was absence of predictable maintenance solutions.

Solution for an advanced fleet management application.

Agile Soft Systems scrutinized the existing system and developed user stories with three user types: fleet managers, drivers, and end customers. As an example, in case of milk transportation, the end customers would be dairy farms and dairy processing units.

Agile methodology

We adopted agile development methodology and began by architecting the system with user stories, persistently collaborating with business stakeholders at every crucial step of the development process. Our aim was to deliver value with each iteration and gather feedback from the stakeholders for improvement as we progressed.

This approach also helped our teams to deliver solutions at a good pace without being overwhelmed by the expanse of the project and its technological challenges. It also helped us to deal with technical debt more prudently as we progressed with the development.

Leveraging Edge Computing

Our aim was to empower the drivers with useful information, while taking enough care we don't overwhelm them with task. We achieved this by designing a smart user experience and by introducing edge computing in our solution.

The edge hardware of choice for this use case was the NVIDIA Jetson Nano. Powered with a powerful GPU, a Quad-core ARM Cortex-A57 MPCore processor, a memory of 4 GB 64-bit LPDDR4, and robust networking capabilities, the edge device was the right fit for our solution and proved to be cost effective for our client's end customers as well.

The edge device was attached and configured with the following: Door, humidity, and temperature sensors, along with cameras that were mounted in their appropriate locations. Besides these, the edge device was also attached to sensors on trucks to enable predictive maintenance and improve their efficiency: engine temperature sensors, oil pressure, fuel consumption, transmission temperature, driveline vibration, brakepad wear sensors and tire pressure sensors were few main ones besides others.

Leading truck manufacturers usually equip trucks with most useful sensors. In case some sensors are missing for a specific task, our client uses universal standard sensors from Continental, Bosch and Honeywell, as preferred by the fleet management companies.

The edge device accesses data from the sensors and is able to notify and alert the driver instantly in regards to the state of the doors — whether they're open or close, cold storage temperature and humidity, route optimization and predictive vehicle maintenance. The low latency and immediate feedback was only possible by adopting edge computing architecture.

Machine Learning

Agile Soft Systems being mindful of the processing power of Jetson Nano, trained a lightweight AI model, MobileNet on Tensorflow and OpenCV to introduce AI and ML abilities in to the mix. The system alerts and notifies truck drivers about the upcoming maintenance of the vehicles, with the help of ML's predictability algorithms.

Our team utilized OpenCV and trained it further to examine video streams from the driver facing cameras in trucks. The system now instantly notifies the driver as soon as it detects drowsiness or cell phone distraction, that has helped improve the safety of drivers to a great extent.

The low latency edge computing solution architected by Agile Soft System, notifies and alerts drivers in real-time, while cueing data transmission to the cloud for further computation to serve fleet managers and end customers.

Conversational AI

Agile Soft Systems implemented Conversational AI to interact with the drivers via a VoiceBot, enhancing the user experience.

We've used Rasa, an open source conversational AI framework and spaCy for NLP tasks such as tokenization, lemmatization, named entity recognition, part-of-speech tagging and dependency parsing. Our team also included Mycroft AI for speech to text and text to speech synthesis.

Business Benefit

The improved fleet management software benefits all user types immeasurably.

With enhanced user experience, it has improved safety for drivers, while reducing cognitive load. The system alerts the drivers immediately by detecting drowsiness and cell phone distraction. It provides real-time information on route optimization, predictive vehicle maintenance, and cold storage temperature and humidity, freeing up the driver to focus on driving. The upgraded system can interact with the drivers via a VoiceBot, making the application easy to use and efficient.

The software provides fleet managers with real-time visibility related to the location and status of all vehicles in the fleet. They are able to optimize routes and reduce downtime due to predictive maintenance, helping them to reduce costs associated with fuel consumption, maintenance, and accidents. The system also logs and reports incidents related to drowsiness and distraction in drivers which helps the managers to take corrective action if required.

The end customers benefit largely by ensuring the safety of perishable goods during transportation by monitoring cold chain conditions. The system helps to reduce costs associated with product spoilage and delivery delays improving customer satisfaction.

Overall, the businesses reported a reduction in travel distance by 3 - 5%, while predictive maintenance helped reduce the unplanned downtime of fleet management companies by 18 to 20%. By enabling continuous monitoring in cold storage trucks, the product spoilage was reported to have reduced by about whopping 40%.

Agile Soft Systems not only addressed the client's critical business and technological challenges but also elevated the safety, efficiency, and overall performance of the entire fleet management process.

Partner with Agile Soft Systems to unlock success and advancement in logistics and revolutionize your business.

Embrace the future of fleet management with us!